docs(agents): Gitea API jobs endpoint takes id, not run_number

GET /api/v1/repos/{owner}/{repo}/actions/runs/{n}/jobs expects the internal run
`id`, not the `run_number` the UI shows as `#239`. The two diverge widely
(id=534 was run_number=238) and the wrong key does NOT error — it silently
returns another run's jobs.

This cost a wrong conclusion while verifying that a push to main no longer
triggers image builds: querying .../runs/238/jobs returned
docs-check/validate-base/validate-omos, which made a lint.yml run look like it
had built images. Corrected by reading `id` from the run listing.

Records the two reliable patterns in the existing "Gitea API access" section:
filter the runs listing on head_sha to learn authoritatively which runs a commit
triggered, then key the jobs endpoint by that id. Also notes that counting runs
per commit is the cheapest assertion of the trigger model (2 runs before the
split, 1 after), and that lint.yml's actionlint job can take 6–15 min because it
apt-installs shellcheck in-container — so a still-running lint is not a problem
signal.

.env.example

@@ -6,8 +6,12 @@
 # Which provider to auto-configure (anthropic, openai, amazon-bedrock)
 OPENCODE_PROVIDER=anthropic
 
-# Model override (optional, defaults per provider)
-# OPENCODE_MODEL=anthropic/claude-sonnet-4-6
+# Model override (optional). Unset = the per-provider default baked into
+# generate-config.py: anthropic/claude-opus-5, amazon-bedrock/
+# global.anthropic.claude-opus-5, or openai/gpt-5.6. Set this to use any other
+# model — the value is written verbatim as the `model` field, so it works for
+# providers with no baked default too. Format: <provider>/<model>.
+# OPENCODE_MODEL=anthropic/claude-opus-5
 
 # ── API Keys (set the one matching your provider) ────────────────────
 # ANTHROPIC_API_KEY=
@@ -31,14 +35,30 @@ WORKSPACE_PATH=~/projects
 # Path to SSH keys on host
 SSH_KEY_PATH=~/.ssh
 
+# ── MemPalace memory (local by default) ───────────────────────────
+# By default each container runs its OWN MemPalace (a local stdio server;
+# palace stored at ~/.mempalace). Uncomment the devbox-palace volume in
+# docker-compose.yml to persist it across container recreation.
+#
+# To instead share ONE MemPalace across several containers / harnesses
+# (pi + opencode + native), point every container at an external HTTP
+# endpoint. When MEMPALACE_REMOTE_URL is set, no local mempalace-mcp is
+# spawned and the devbox-palace volume is irrelevant.
+#   Serve it with:  mempalace-mcp --transport http --host 0.0.0.0 --port 8765
+# MEMPALACE_REMOTE_URL=http://mempalace.lan:8765/mcp
+# MEMPALACE_REMOTE_TOKEN=          # optional — sent as: Authorization: Bearer <token>
+
 # ── LAN access from the container (host-OS-agnostic) ─────────────────
 # On VM-backed hosts (macOS OrbStack / Docker Desktop, also Docker Desktop
 # on Windows) the container runs in a Linux VM and CANNOT reach the host's
 # directly-attached LAN peers by default. On native Linux Docker the LAN is
 # reachable directly and nothing is needed. The entrypoint detects this and,
 # on VM-backed hosts, generates ~/.ssh-local/config so the host can be used
-# as an SSH jump (use the `dssh` alias, or add `ProxyJump host` to targets
-# in your bind-mounted ~/.ssh/config).
+# as an SSH jump (use the `dssh` alias). Reach the host itself with
+# `dssh host`. To reach named LAN peers, put `ProxyJump host` overrides in a
+# host-owned ~/.config/devbox-shell/ssh-lan.conf (bind-mounted in) rather than
+# editing your ~/.ssh/config — see ssh-lan.conf.example. Public-IP hosts (and
+# anything reached via a public jump host) connect directly, no jump needed.
 #
 # DEVBOX_LAN_ACCESS: auto (default) | jump | off
 #   auto = set up the jump only on VM-backed hosts; no-op on native Linux.
@@ -54,6 +74,12 @@ SSH_KEY_PATH=~/.ssh
 #
 # DEVBOX_HOST_ALIAS: host hostname to reach (default host.docker.internal).
 # DEVBOX_HOST_ALIAS=host.docker.internal
+#
+# DEVBOX_LAN_AUTOJUMP_PRIVATE: 1 = ProxyJump ANY RFC1918 (private) IP through
+# the host, so bare `dssh user@<ip>` works on whatever LAN the (roaming) host
+# is currently joined to, without naming peers. Matches the typed address, not
+# the resolved HostName, so named hosts with their own ProxyJump are unaffected.
+# DEVBOX_LAN_AUTOJUMP_PRIVATE=0
 
 # ── Skillset (agent skills and instructions) ─────────────────────────
 # If you have a skillset repo, the entrypoint auto-deploys skills and
@@ -89,34 +115,15 @@ SSH_KEY_PATH=~/.ssh
 # Requires image built with INSTALL_OMOS=true
 # ENABLE_OMOS=false
 # OMOS_TMUX=false              # Enable tmux multiplexer integration
-# OMOS_SKILLS=true             # Install recommended skills (simplify, agent-browser, cartography)
-# OMOS_RESET=false             # Force regenerate oh-my-opencode-slim config on next start
+# OMOS_SKILLS=true             # Symlink bundled OMOS skills (clonedeps, codemap,
+#                              # deepwork, oh-my-opencode-slim, simplify) from the
+#                              # image into ~/.agents/skills/ each start; updates
+#                              # on image pull. Independent of ENABLE_OMOS.
+#                              # See docs/omos-skills.md
+# OMOS_RESET=false             # Force regenerate oh-my-opencode-slim config on next start (does not affect skills)
 
-# ── pi coding-agent (alternative/complementary harness) ─────────────────
-# Requires image built with INSTALL_PI=true.
-# When the image is built with both INSTALL_OPENCODE=true (default) and
-# INSTALL_PI=true, both harnesses share the same mempalace install and
-# palace path — wing data is mutually visible to either harness.
-#
-# Pi version is baked at build time via PI_VERSION (default: latest at
-# build). The baked `pi` binary is at /usr/bin/pi (system npm prefix);
-# rebuild the image to upgrade it. NPM_CONFIG_PREFIX is set to
-# /home/developer/.pi/npm-global, so anything installed via
-# `pi install npm:...` or `npm install -g` as the developer user
-# (themes, skills, extensions, including a user-installed pi itself)
-# lands on the named volume and survives container recreate AND image
-# rebuilds. A user-installed pi wins via PATH order over the baked one.
-#
-# Pi config (settings.json, extensions toggle state, sessions, auth) persists in the
-# devbox-pi-config named volume mounted at ~/.pi/.
-#
-# To launch pi from a `compose run` invocation:
-#   docker compose run --rm devbox pi
-# To attach to a running container:
-#   docker compose exec -u developer devbox pi
-# Default `compose run` (no args) drops to bash; pick the harness yourself.
-#
-# Build args (set in docker-compose.yml or via --build-arg on docker build):
-#   INSTALL_PI=true              # default false; opt-in
-#   PI_VERSION=latest            # pin a specific version, e.g. 0.73.0
-#   INSTALL_OPENCODE=false       # build a pi-only image (still has Bun in -omos)
+# ── Background subagents (opencode experimental flag) ────────────────
+# Baked ON in the image (ENV in Dockerfile.base) because OMOS V2+ default
+# orchestration depends on opencode's native background subagents. Uncomment
+# to opt out (e.g. to force blocking/foreground orchestration).
+# OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=false

.env.shared.example

@@ -27,6 +27,13 @@ GIT_USER_EMAIL=your.name@example.com
 # If you have per-user keys: SSH_KEY_PATH=~/<signum>/.ssh
 # SSH_KEY_PATH=~/.ssh
 
+# ── MemPalace memory ─────────────────────────────────────────────────
+# Default: each container runs its own LOCAL palace. To share ONE external
+# MemPalace across containers/harnesses, set the URL (no local server is
+# spawned when set). MEMPALACE_REMOTE_TOKEN is an optional bearer token.
+# MEMPALACE_REMOTE_URL=http://mempalace.lan:8765/mcp
+# MEMPALACE_REMOTE_TOKEN=
+
 # ── Locale (defaults to en_US.UTF-8) ────────────────────────────────
 # LANG=sv_SE.UTF-8
 # LANGUAGE=sv_SE:sv

.gitea/README.md

@@ -8,14 +8,16 @@ the build pipeline is shaped the way it is, you're in the right place.
 
 | File | Trigger | Role |
 |---|---|---|
-| [`workflows/docker-publish-split.yml`](workflows/docker-publish-split.yml) | `push: tags: v*` | **Production release pipeline.** Two-phase split-base build: shared `base-<hash>` published once (skipped on cache hit), then five parallel variant deltas. ~40–80 min wall clock depending on runner count and whether base needs rebuilding. |
-| [`workflows/validate.yml`](workflows/validate.yml) | `push: branches: main` + PR | **Lightweight gate.** amd64-only smoke test of all five variants + `DOCKER_HUB.md` sync check. ~30 min. Fires on every push to `main`. |
+| [`workflows/docker-publish-split.yml`](workflows/docker-publish-split.yml) | `push: tags: v*` | **Production release pipeline.** Two-phase split-base build: shared `base-<hash>` published once (skipped on cache hit), then two parallel variant deltas. ~40–80 min wall clock depending on runner count and whether base needs rebuilding. |
+| [`workflows/validate.yml`](workflows/validate.yml) | `push: branches: main` + PR | **Lightweight gate.** amd64-only smoke test of both variants + `DOCKER_HUB.md` sync check. ~30 min. Fires on every push to `main`. |
 
 ## Why the split-base pipeline exists
 
-opencode-devbox builds **five image variants** (`base`, `omos`, `with-pi`, `omos-with-pi`, `pi-only`) × **two architectures** (amd64, arm64). Four opencode-bearing variants publish under this repo (**eight tags per release** + the floating `base-latest`); the `pi-only` build is pushed into the separate `joakimp/pi-devbox` repo as `base-pi-only` (so no opencode-less tag appears here). Today's runners are 2 self-hosted gitea Actions runners. arm64 builds are emulated under QEMU, which is the dominant cost (~3–5x slower than native).
+opencode-devbox builds **two image variants** (`base`, `omos`) × **two architectures** (amd64, arm64), publishing **four tags per release** + the floating `base-latest`. Today's runners are 2 self-hosted gitea Actions runners. arm64 builds are emulated under QEMU, which is the dominant cost (~3–5x slower than native).
 
-The five variants share ~95% of their layers (Debian + apt + Node + AWS CLI + mempalace + dev tools + entrypoints). The original `Dockerfile` was a single multi-stage build with `INSTALL_*` build-args gating variant-specific RUNs. BuildKit's per-layer cache key is content-addressed, but as soon as a build-arg-gated `RUN` produces a different layer hash for variant A vs variant B, every subsequent layer also has a different parent → identical commands re-execute per variant. Result: minimal cross-variant cache reuse on a fresh build.
+> pi was removed in v2.0.0; it now builds in its own `joakimp/pi-devbox` repo. Before v2.0.0 a fifth `pi-only` build was produced here and pushed into that repo as `base-pi-only` — that coupling is gone.
+
+The two variants share ~95% of their layers (Debian + apt + Node + AWS CLI + mempalace + dev tools + entrypoints). The original `Dockerfile` was a single multi-stage build with `INSTALL_*` build-args gating variant-specific RUNs. BuildKit's per-layer cache key is content-addressed, but as soon as a build-arg-gated `RUN` produces a different layer hash for variant A vs variant B, every subsequent layer also has a different parent → identical commands re-execute per variant. Result: minimal cross-variant cache reuse on a fresh build.
 
 Two improvements were considered:
 
@@ -32,8 +34,8 @@ The split-base architecture is what the `docker-publish-split.yml` workflow exer
                        │                  │   probe Docker Hub.
                        │  hash inputs:    │   (resolve-versions
                        │   Dockerfile.base│   runs in parallel:
-                       │   rootfs/        │   npm view pi/omos
-                       │   entrypoint*.sh │   → concrete versions)
+                       │   rootfs/        │   npm view omos
+                       │   entrypoint*.sh │   → concrete version)
                        └────────┬─────────┘
                                 │
                   ┌─────────────┴─────────────┐
@@ -47,18 +49,18 @@ The split-base architecture is what the `docker-publish-split.yml` workflow exer
                        └────────┬─────────┘   to Docker Hub.
                                 │
         ┌───────────────────────┼───────────────────────┐
-        ▼                       ▼                       ▼
-   ┌──────────┐            ┌──────────┐         ┌──────────────┐
-   │smoke-base│            │smoke-omos│   ...   │smoke-omos-pi │   amd64 only,
-   └────┬─────┘            └────┬─────┘         └──────┬───────┘   parallel.
-        │                       │                      │
-        ▼                       ▼                      ▼
-   ┌──────────┐            ┌──────────┐         ┌──────────────┐
-   │build-    │            │build-    │         │build-        │   multi-arch,
-   │variant-  │            │variant-  │   ...   │variant-      │   parallel,
-   │base      │            │omos      │         │omos-with-pi  │   tag push.
-   └────┬─────┘            └────┬─────┘         └──────┬───────┘
-        └───────────────────────┴──────────────────────┘
+        ▼                       ▼
+   ┌──────────┐            ┌──────────┐
+   │smoke-base│            │smoke-omos│             amd64 only,
+   └────┬─────┘            └────┬─────┘             parallel.
+        │                       │
+        ▼                       ▼
+   ┌──────────┐            ┌──────────┐
+   │build-    │            │build-    │             multi-arch,
+   │variant-  │            │variant-  │             parallel,
+   │base      │            │omos      │             tag push.
+   └────┬─────┘            └────┬─────┘
+        └───────────┬───────────┘
                                 │
                                 ▼
                   ┌──────────────────────────┐
@@ -73,7 +75,13 @@ The split-base architecture is what the `docker-publish-split.yml` workflow exer
                   └──────────────────────────┘
 ```
 
-### Step 1: `base-decide` (and `resolve-versions` in parallel)
+### Step 1: `resolve-versions`, then `base-decide`
+
+**`resolve-versions`** resolves floating refs to concrete values: `omos_version`
+(npm `latest`) and `mempalace_toolkit_ref` (the `mempalace-toolkit` `main` HEAD
+resolved to a commit SHA via the gitea commits API). **`base-decide`** now
+**depends on `resolve-versions`** (they no longer run in parallel) because it
+folds `mempalace_toolkit_ref` into the base hash — see below.
 
 **`base-decide`** computes a SHA-256 hash over the inputs that determine
 the base image's content:
@@ -88,6 +96,9 @@ the base image's content:
     ! -name '._*' \
     -print0 | sort -z | xargs -0 cat
   cat entrypoint.sh entrypoint-user.sh
+  echo "$mempalace_toolkit_ref"   # CI-resolved SHA; mempalace-toolkit is
+                                  # cloned in Dockerfile.base, so a moved
+                                  # toolkit must force a base rebuild
 } | sha256sum | cut -c1-12
 ```
 
@@ -111,13 +122,12 @@ dependency between them) and resolves the floating npm packages whose
 `*_VERSION` build-args default to `latest`:
 
 ```sh
-PI_VERSION=$(npm view @earendil-works/pi-coding-agent version)
 OMOS_VERSION=$(npm view oh-my-opencode-slim version)
 ```
 
-The outputs (`pi_version`, `omos_version`) are consumed by every variant
-smoke and build job that installs pi or omos. **Why this exists:** without
-it, the `npm install -g` RUN layer in `Dockerfile.variant` hashes
+The output (`omos_version`) is consumed by the omos variant smoke and
+build jobs. **Why this exists:** without it, the `npm install -g` RUN
+layer in `Dockerfile.variant` hashes
 identically across builds (same ARG default, same command string), so
 the registry buildcache silently reuses the layer from whatever upstream
 version was current when the cache was first populated. This is the
@@ -125,9 +135,9 @@ cache-hit silent-regression class of bug that shipped pi-devbox v0.74.0
 through v0.75.5 with identical image bytes (fixed in pi-devbox v0.75.5b
 2026-05-23). Currently masked here by `OPENCODE_VERSION` bumping every
 release (parent-chain cache-key invalidation), but masking would fail on
-a `vN.N.Nb` opencode-version-unchanged release that only bumps pi or
-omos. Smoke jobs additionally assert `EXPECTED_PI_VERSION` /
-`EXPECTED_OMOS_VERSION` against the resolved values.
+a `vN.N.Nb` opencode-version-unchanged release that only bumps omos.
+Smoke jobs additionally assert `EXPECTED_OMOS_VERSION` against the
+resolved value.
 
 ### Step 2: `build-base` (conditional)
 
@@ -139,23 +149,21 @@ when only one or two layers changed.
 The base image is **not** tagged `base-latest` here — that promotion
 happens at the very end after all variants succeed (see step 5).
 
-### Step 3: `smoke-*` (×4, parallel)
+### Step 3: `smoke-*` (×2, parallel)
 
 For each variant: build amd64-only against the base tag, load into
 local docker, run [`scripts/smoke-test.sh`](../scripts/smoke-test.sh).
 Variant build-args:
 
-| variant | INSTALL_OPENCODE | INSTALL_OMOS | INSTALL_PI |
-|---|---|---|---|
-| `base` | true | false | false |
-| `omos` | true | true | false |
-| `with-pi` | true | false | true |
-| `omos-with-pi` | true | true | true |
+| variant | INSTALL_OPENCODE | INSTALL_OMOS |
+|---|---|---|
+| `base` | true | false |
+| `omos` | true | true |
 
 Smoke runs `--variant <name>` to enable variant-specific assertions.
 Gate the publish: a smoke failure for variant X blocks `build-variant-X`.
 
-### Step 4: `build-variant-*` (×4, parallel)
+### Step 4: `build-variant-*` (×2, parallel)
 
 For each variant that passed smoke: multi-arch (amd64 + arm64) build of
 `Dockerfile.variant`, pushed to Docker Hub with the user-facing release
@@ -165,8 +173,6 @@ tags:
 |---|---|
 | `build-variant-base` | `vX.Y.Z`, `latest` |
 | `build-variant-omos` | `vX.Y.Z-omos`, `latest-omos` |
-| `build-variant-with-pi` | `vX.Y.Z-with-pi`, `latest-with-pi` |
-| `build-variant-omos-with-pi` | `vX.Y.Z-omos-with-pi`, `latest-omos-with-pi` |
 
 The `latest*` aliases are only updated when `promote_latest=true` (the
 manual dispatch input) — for test runs, `promote_latest=false` keeps the
@@ -174,7 +180,7 @@ production aliases pointing at the previous good release.
 
 ### Step 5: `promote-base-latest`
 
-Once all five variants successfully publish, re-tag `base-<hash>` as
+Once both variants successfully publish, re-tag `base-<hash>` as
 `base-latest` using `crane copy`. This is a **manifest-level re-tag, not
 a rebuild** — it touches only Docker Hub's image index, takes seconds,
 and is atomic.
@@ -182,7 +188,7 @@ and is atomic.
 The reason this happens *after* variants succeed (rather than alongside
 `build-base`) is so a partial failure leaves `base-latest` pointing at
 the previous known-good base. External consumers who pin to
-`base-latest` (e.g. the planned pi-devbox repo) never see a broken base.
+`base-latest` never see a broken base.
 
 ### Step 6: `update-description`
 
@@ -194,18 +200,21 @@ field via the Hub REST API. Same step as the production pipeline.
 The base sets
 
 ```
-ENV NPM_CONFIG_PREFIX=/home/developer/.pi/npm-global
+ENV NPM_CONFIG_PREFIX=/home/developer/.config/opencode/npm-global
 ```
 
-This is intentional — it makes `pi install npm:<pkg>` and `npm install -g`
-land on the `devbox-pi-config` named volume at runtime, so user-installed
-packages survive container recreate AND image rebuild.
+This is intentional — it makes `npm install -g` land on the
+`devbox-opencode-config` named volume at runtime, so user-installed
+packages survive container recreate AND image rebuild. (Before v2.0.0
+this prefix lived at `~/.pi/npm-global` on the now-removed
+`devbox-pi-config` volume; `entrypoint-user.sh` migrates the old path
+once.)
 
 But the *variant build* inherits this prefix at build time. If left as-is,
 `npm install -g opencode-ai@$VERSION` in `Dockerfile.variant` would
-install opencode into `/home/developer/.pi/npm-global/...`, which is then
-**shadowed by the volume mount at runtime** → opencode disappears from
-PATH on first start.
+install opencode into `/home/developer/.config/opencode/npm-global/...`,
+which is then **shadowed by the volume mount at runtime** → opencode
+disappears from PATH on first start.
 
 Fix: each `npm install -g` in `Dockerfile.variant` overrides the prefix
 per-RUN:
@@ -216,7 +225,7 @@ RUN NPM_CONFIG_PREFIX=/usr npm install -g opencode-ai@${OPENCODE_VERSION}
 
 Baked binaries land on `/usr/bin/...` (system prefix), survive the volume
 mount. Runtime-installed user packages still land on
-`~/.pi/npm-global/...`. Both visible on PATH.
+`~/.config/opencode/npm-global/...`. Both visible on PATH.
 
 ## Cache strategy
 
@@ -238,7 +247,7 @@ matters more.
 
 | Scenario | Production pipeline | Split-base pipeline |
 |---|---|---|
-| Version-bump-only release (only opencode/pi/omos version changed) | ~165–180 min | **~30–40 min** (base cache hit) |
+| Version-bump-only release (only opencode/omos version changed) | ~165–180 min | **~30–40 min** (base cache hit) |
 | Base-touching release (apt/Node/Debian/entrypoint change) | ~165–180 min | **~70–90 min** (base rebuilds) |
 
 The split-base pipeline pays its dues on base-touching releases (which are
@@ -252,7 +261,7 @@ on every push to `main` and on PRs. It:
 
 1. Runs `scripts/generate-dockerhub-md.py --check` to enforce
    `DOCKER_HUB.md` is in sync with `HUB_TEMPLATE`.
-2. Builds each of the five variants amd64-only (no multi-arch, no push)
+2. Builds each of the two variants amd64-only (no multi-arch, no push)
    and runs `scripts/smoke-test.sh`.
 
 This catches regressions before they reach a tag push. Wall clock ~30 min.

.gitea/workflows/docker-publish-split.yml

@@ -7,10 +7,10 @@ name: Publish Docker Image
 #   1. base-decide      compute base hash from Dockerfile.base + rootfs/
 #                       + entrypoints; probe Docker Hub for existing tag.
 #   2. build-base       only if probe missed; multi-arch push of base-<hash>.
-#   3. smoke-* (×4)     amd64-only build of each variant FROMing the base
+#   3. smoke-* (×2)     amd64-only build of each variant FROMing the base
 #                       tag; runs scripts/smoke-test.sh.
 #   4. build-variant-*  multi-arch push of each variant tag (the user-
-#       (×4)            facing release tags, unchanged in shape).
+#       (×2)            facing release tags, unchanged in shape).
 #   5. promote-base-latest  re-tag base-<hash> → base-latest with `crane copy`
 #                       (manifest copy, no rebuild).
 #   6. update-description   patch Docker Hub description (unchanged).
@@ -34,14 +34,20 @@ concurrency:
   group: ${{ github.workflow }}-${{ github.ref }}
   cancel-in-progress: false
 
+# Gitea Actions' default step shell is `sh -e {0}` (dash), which rejects
+# bash-only syntax like `set -o pipefail`, `[[ ]]`, and arrays. Setting the
+# default to bash workflow-wide eliminates the whole class of "forgot
+# `shell: bash` on this step" bugs. (Ported from pi-devbox, where this class
+# bit twice: ed49b8d resolve-versions, b7197e8/b33e9dc promote-base-latest,
+# run 418.) All existing dash steps use only POSIX syntax, so bash (a
+# superset) runs them unchanged. Enforced by lint.yml's shell guard.
+defaults:
+  run:
+    shell: bash
+
 env:
   BUILDKIT_PROGRESS: plain
   IMAGE: ${{ vars.DOCKERHUB_USERNAME }}/opencode-devbox
-  # The pi-only variant is built here (single source of truth for the pi stack)
-  # but published into the pi-devbox repo as an internal building-block tag,
-  # NOT under opencode-devbox — so opencode-devbox never shows a tag with no
-  # opencode in it. pi-devbox's own CI FROMs PI_IMAGE:base-pi-only.
-  PI_IMAGE: ${{ vars.DOCKERHUB_USERNAME }}/pi-devbox
   RELEASE_TAG: ${{ github.ref_type == 'tag' && github.ref_name || inputs.release_tag }}
   PROMOTE_LATEST: ${{ github.ref_type == 'tag' && 'true' || inputs.promote_latest }}
 
@@ -53,6 +59,7 @@ env:
 jobs:
   # ── Phase 1: decide whether base needs rebuilding ──────────────────
   base-decide:
+    needs: [resolve-versions]
     runs-on: ubuntu-latest
     container:
       image: catthehacker/ubuntu:act-latest
@@ -63,6 +70,9 @@ jobs:
       - name: Checkout
         uses: actions/checkout@v4
 
+      - name: Guard — base *_REF args must be folded into the base hash
+        run: bash scripts/check-base-hash.sh
+
       - name: Compute base tag from Dockerfile.base + dependencies
         id: compute
         run: |
@@ -82,6 +92,10 @@ jobs:
                 ! -name '._*' \
                 -print0 2>/dev/null | sort -z | xargs -0 cat 2>/dev/null
               cat entrypoint.sh entrypoint-user.sh
+              # mempalace-toolkit is cloned in Dockerfile.base at a ref CI
+              # resolves to a SHA; fold it in so base_tag changes when the
+              # toolkit moves (otherwise a toolkit-only fix never lands).
+              echo "${{ needs.resolve-versions.outputs.mempalace_toolkit_ref }}"
             } | sha256sum | cut -c1-12
           )
           BASE_TAG="base-${HASH}"
@@ -107,63 +121,75 @@ jobs:
             echo "Base tag ${IMAGE}:${{ steps.compute.outputs.base_tag }} missing — will build."
           fi
 
-  # ── Phase 1b: resolve floating npm versions (pi, omos) to concrete
+  # ── Phase 1b: resolve floating npm versions (omos) to concrete
   # versions so the variant build-args carry a different value when an
-  # upstream package bumps. Without this, when PI_VERSION / OMOS_VERSION
-  # default to 'latest', the docker/build-push-action build-arg string
-  # is byte-identical across builds, so the resulting layer-hash is
-  # identical, so the registry buildcache silently reuses the layer
-  # from whatever pi/omos version was current when the cache was first
-  # populated. Same class of bug as pi-devbox v0.74.0..v0.75.5 (fixed in
-  # v0.75.5b 2026-05-23). Currently masked here because OPENCODE_VERSION
-  # is hard-coded in Dockerfile.variant and bumps every release —
-  # invalidating the parent-chain cache key for the pi/omos layers — but
-  # that masking would fail the moment we cut a vN.N.Nb opencode-version-
-  # unchanged release that only bumps pi or omos. Fix is preventative.
+  # upstream package bumps. Without this, when OMOS_VERSION defaults to
+  # 'latest', the docker/build-push-action build-arg string is byte-
+  # identical across builds, so the resulting layer-hash is identical,
+  # so the registry buildcache silently reuses the layer from whatever
+  # omos version was current when the cache was first populated. Same
+  # class of bug as pi-devbox v0.74.0..v0.75.5 (fixed in v0.75.5b
+  # 2026-05-23). Currently masked because OPENCODE_VERSION is hard-coded
+  # in Dockerfile.variant and bumps every release — invalidating the
+  # parent-chain cache key for the omos layer — but that masking would
+  # fail the moment we cut a vN.N.Nb opencode-version-unchanged release
+  # that only bumps omos. Fix is preventative.
   resolve-versions:
     runs-on: ubuntu-latest
     container:
       image: catthehacker/ubuntu:act-latest
     outputs:
-      pi_version: ${{ steps.resolve.outputs.pi_version }}
       omos_version: ${{ steps.resolve.outputs.omos_version }}
-      fork_ref: ${{ steps.resolve.outputs.fork_ref }}
-      obsmem_ref: ${{ steps.resolve.outputs.obsmem_ref }}
+      mempalace_toolkit_ref: ${{ steps.resolve.outputs.mempalace_toolkit_ref }}
     steps:
-      - name: Resolve pi + omos versions from npm registry
+      - name: Resolve omos version from npm registry
         id: resolve
+        shell: bash
         run: |
-          set -eu
+          set -euo pipefail
+          # Fail loud rather than silently shipping a floating ref or a bad
+          # version. A transient network/API failure must ABORT the release,
+          # not bake an unpinned ref that defeats both cache-busting AND
+          # after-the-fact reproducibility. (Previously the gitea lookup fell
+          # back to `main` via `|| echo`, and the npm lookup had no guard.)
+          # NOTE: shell: bash is REQUIRED — `set -o pipefail` is illegal in
+          # the runner's default dash/sh and aborts the step immediately.
+          require_sha() { # $1=label $2=value
+            if ! printf '%s' "${2:-}" | grep -qiE '^[0-9a-f]{40}$'; then
+              echo "::error::Could not resolve $1 to a commit SHA (got '${2:-<empty>}'). Refusing to fall back to a floating ref — published images must stay reproducible. Check connectivity and GITEA_BUILD_TOKEN/GITHUB_TOKEN."
+              exit 1
+            fi
+          }
           # Query the npm registry directly via curl+jq rather than `npm view`.
           # catthehacker/ubuntu:act-latest ships Node/npm under /opt/acttoolcache/
           # and adds it to PATH only via /etc/environment — which act_runner never
           # sources (it reads the Docker image's ENV instructions, not /etc/environment).
           # curl and jq are both guaranteed present in every job in this workflow.
-          PI_VERSION=$(curl -sf "https://registry.npmjs.org/@earendil-works%2Fpi-coding-agent/latest" | jq -r '.version')
-          OMOS_VERSION=$(curl -sf "https://registry.npmjs.org/oh-my-opencode-slim/latest" | jq -r '.version')
-          echo "pi_version=${PI_VERSION}"     >> "$GITHUB_OUTPUT"
+          OMOS_VERSION=$(curl -sf "https://registry.npmjs.org/oh-my-opencode-slim/latest" | jq -r '.version' 2>/dev/null || true)
+          if ! printf '%s' "${OMOS_VERSION:-}" | grep -qE '^[0-9]+\.[0-9]+\.[0-9]+'; then
+            echo "::error::Could not resolve oh-my-opencode-slim version from npm (got '${OMOS_VERSION:-<empty>}'). Refusing to build with an unresolved version."
+            exit 1
+          fi
           echo "omos_version=${OMOS_VERSION}" >> "$GITHUB_OUTPUT"
-          # Resolve the pi-fork / pi-observational-memory git refs (default
-          # branch master) to concrete commit SHAs so the build-arg string
-          # changes whenever upstream moves — defeating the same registry-
-          # buildcache cache-hit footgun that PI_VERSION/OMOS_VERSION guard
-          # against. The Accept: application/vnd.github.sha media type returns
-          # the bare SHA. Falls back to the branch name if the API is
-          # unreachable/rate-limited (still functional, just cache-stale-prone).
-          FORK_REF=$(curl -sf -H "Accept: application/vnd.github.sha" \
-            "https://api.github.com/repos/elpapi42/pi-fork/commits/master" || echo "master")
-          OBSMEM_REF=$(curl -sf -H "Accept: application/vnd.github.sha" \
-            "https://api.github.com/repos/elpapi42/pi-observational-memory/commits/master" || echo "master")
-          [ -n "$FORK_REF" ]   || FORK_REF=master
-          [ -n "$OBSMEM_REF" ] || OBSMEM_REF=master
-          echo "fork_ref=${FORK_REF}"     >> "$GITHUB_OUTPUT"
-          echo "obsmem_ref=${OBSMEM_REF}" >> "$GITHUB_OUTPUT"
-          echo "Resolved PI_VERSION=${PI_VERSION}, OMOS_VERSION=${OMOS_VERSION}"
-          echo "Resolved PI_FORK_REF=${FORK_REF}, PI_OBSMEM_REF=${OBSMEM_REF}"
+          echo "Resolved OMOS_VERSION=${OMOS_VERSION}"
+          # Resolve mempalace-toolkit main HEAD to a commit SHA. Unlike omos
+          # (an npm pkg baked into the VARIANT), mempalace-toolkit is cloned
+          # in Dockerfile.base, so this SHA is ALSO folded into the
+          # base-decide hash to force a base rebuild when the toolkit moves
+          # (without it, a toolkit-only fix silently fails to land unless
+          # Dockerfile.base itself changes). gitea allows unauthenticated
+          # public-repo commit listing; the token header is harmless if the
+          # env vars are unset (degrades to anon, still HTTP 200).
+          MEMPALACE_TOOLKIT_REF=$(curl -sf -H "Authorization: token ${GITEA_BUILD_TOKEN:-${GITHUB_TOKEN:-}}" \
+            "https://gitea.jordbo.se/api/v1/repos/joakimp/mempalace-toolkit/commits?limit=1&sha=main" \
+            | jq -r '.[0].sha // empty' 2>/dev/null || true)
+          require_sha MEMPALACE_TOOLKIT_REF "$MEMPALACE_TOOLKIT_REF"
+          echo "mempalace_toolkit_ref=${MEMPALACE_TOOLKIT_REF}" >> "$GITHUB_OUTPUT"
+          echo "Resolved MEMPALACE_TOOLKIT_REF=${MEMPALACE_TOOLKIT_REF}"
 
   # ── Phase 2: build & push base (multi-arch), only when needed ──────
   build-base:
-    needs: [base-decide]
+    needs: [base-decide, resolve-versions]
     if: needs.base-decide.outputs.need_build == 'true'
     runs-on: ubuntu-latest
     container:
@@ -211,6 +237,7 @@ jobs:
         shell: bash
         env:
           BASE_TAG_FULL: ${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
+          MEMPALACE_TOOLKIT_REF: ${{ needs.resolve-versions.outputs.mempalace_toolkit_ref }}
         run: |
           set -euo pipefail
           # 3-attempt retry around `docker buildx build --push` for transient
@@ -231,6 +258,7 @@ jobs:
             if docker buildx build \
                 --platform linux/amd64,linux/arm64 \
                 --file Dockerfile.base \
+                --build-arg MEMPALACE_TOOLKIT_REF="${MEMPALACE_TOOLKIT_REF}" \
                 --push \
                 --tag "${BASE_TAG_FULL}" \
                 .; then
@@ -292,7 +320,8 @@ jobs:
             BASE_IMAGE=${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
             INSTALL_OPENCODE=true
             INSTALL_OMOS=false
-            INSTALL_PI=false
+            RELEASE_TAG=smoke
+            SOURCE_REVISION=${{ github.sha }}
       - name: Smoke test (amd64)
         run: bash scripts/smoke-test.sh opencode-devbox:smoke-base --variant base
 
@@ -335,159 +364,17 @@ jobs:
             BASE_IMAGE=${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
             INSTALL_OPENCODE=true
             INSTALL_OMOS=true
-            INSTALL_PI=false
             OMOS_VERSION=${{ needs.resolve-versions.outputs.omos_version }}
+            RELEASE_TAG=smoke
+            SOURCE_REVISION=${{ github.sha }}
       - env:
           EXPECTED_OMOS_VERSION: ${{ needs.resolve-versions.outputs.omos_version }}
         run: bash scripts/smoke-test.sh opencode-devbox:smoke-omos --variant omos
 
-  smoke-with-pi:
-    needs: [base-decide, build-base, resolve-versions]
-    if: |
-      always() &&
-      needs.base-decide.result == 'success' &&
-      needs.resolve-versions.result == 'success' &&
-      (needs.build-base.result == 'success' || needs.build-base.result == 'skipped')
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - uses: actions/checkout@v4
-      - run: echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-      - run: |
-          rm -rf /opt/hostedtoolcache /opt/microsoft /opt/az /opt/ghc \
-            /usr/local/.ghcup /usr/share/dotnet /usr/share/swift \
-            /usr/local/lib/android /usr/local/share/powershell \
-            /usr/local/share/chromium /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-      - uses: docker/setup-buildx-action@v4
-        with: {driver-opts: network=host}
-      - uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-      - uses: docker/build-push-action@v7
-        with:
-          context: .
-          file: Dockerfile.variant
-          platforms: linux/amd64
-          push: false
-          load: true
-          tags: opencode-devbox:smoke-with-pi
-          build-args: |
-            BASE_IMAGE=${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
-            INSTALL_OPENCODE=true
-            INSTALL_OMOS=false
-            INSTALL_PI=true
-            PI_VERSION=${{ needs.resolve-versions.outputs.pi_version }}
-            PI_FORK_REF=${{ needs.resolve-versions.outputs.fork_ref }}
-            PI_OBSMEM_REF=${{ needs.resolve-versions.outputs.obsmem_ref }}
-      - env:
-          EXPECTED_PI_VERSION: ${{ needs.resolve-versions.outputs.pi_version }}
-        run: bash scripts/smoke-test.sh opencode-devbox:smoke-with-pi --variant with-pi
-
-  smoke-omos-with-pi:
-    needs: [base-decide, build-base, resolve-versions]
-    if: |
-      always() &&
-      needs.base-decide.result == 'success' &&
-      needs.resolve-versions.result == 'success' &&
-      (needs.build-base.result == 'success' || needs.build-base.result == 'skipped')
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - uses: actions/checkout@v4
-      - run: echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-      - run: |
-          rm -rf /opt/hostedtoolcache /opt/microsoft /opt/az /opt/ghc \
-            /usr/local/.ghcup /usr/share/dotnet /usr/share/swift \
-            /usr/local/lib/android /usr/local/share/powershell \
-            /usr/local/share/chromium /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-      - uses: docker/setup-buildx-action@v4
-        with: {driver-opts: network=host}
-      - uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-      - uses: docker/build-push-action@v7
-        with:
-          context: .
-          file: Dockerfile.variant
-          platforms: linux/amd64
-          push: false
-          load: true
-          tags: opencode-devbox:smoke-omos-with-pi
-          build-args: |
-            BASE_IMAGE=${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
-            INSTALL_OPENCODE=true
-            INSTALL_OMOS=true
-            INSTALL_PI=true
-            PI_VERSION=${{ needs.resolve-versions.outputs.pi_version }}
-            OMOS_VERSION=${{ needs.resolve-versions.outputs.omos_version }}
-            PI_FORK_REF=${{ needs.resolve-versions.outputs.fork_ref }}
-            PI_OBSMEM_REF=${{ needs.resolve-versions.outputs.obsmem_ref }}
-      - env:
-          EXPECTED_PI_VERSION: ${{ needs.resolve-versions.outputs.pi_version }}
-          EXPECTED_OMOS_VERSION: ${{ needs.resolve-versions.outputs.omos_version }}
-        run: bash scripts/smoke-test.sh opencode-devbox:smoke-omos-with-pi --variant omos-with-pi
-
-  smoke-pi-only:
-    needs: [base-decide, build-base, resolve-versions]
-    if: |
-      always() &&
-      needs.base-decide.result == 'success' &&
-      needs.resolve-versions.result == 'success' &&
-      (needs.build-base.result == 'success' || needs.build-base.result == 'skipped')
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - uses: actions/checkout@v4
-      - run: echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-      - run: |
-          rm -rf /opt/hostedtoolcache /opt/microsoft /opt/az /opt/ghc \
-            /usr/local/.ghcup /usr/share/dotnet /usr/share/swift \
-            /usr/local/lib/android /usr/local/share/powershell \
-            /usr/local/share/chromium /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-      - uses: docker/setup-buildx-action@v4
-        with: {driver-opts: network=host}
-      - uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-      - uses: docker/build-push-action@v7
-        with:
-          context: .
-          file: Dockerfile.variant
-          platforms: linux/amd64
-          push: false
-          load: true
-          tags: opencode-devbox:smoke-pi-only
-          build-args: |
-            BASE_IMAGE=${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
-            INSTALL_OPENCODE=false
-            INSTALL_OMOS=false
-            INSTALL_PI=true
-            PI_VERSION=${{ needs.resolve-versions.outputs.pi_version }}
-            PI_FORK_REF=${{ needs.resolve-versions.outputs.fork_ref }}
-            PI_OBSMEM_REF=${{ needs.resolve-versions.outputs.obsmem_ref }}
-      - env:
-          EXPECTED_PI_VERSION: ${{ needs.resolve-versions.outputs.pi_version }}
-        run: bash scripts/smoke-test.sh opencode-devbox:smoke-pi-only --variant pi-only
-
   # ── Phase 4: multi-arch publish per variant ────────────────────────
 
   build-variant-base:
-    needs: [base-decide, smoke-base]
+    needs: [base-decide, smoke-base, resolve-versions]
     runs-on: ubuntu-latest
     container:
       image: catthehacker/ubuntu:act-latest
@@ -526,8 +413,10 @@ jobs:
         env:
           TAGS: ${{ steps.tags.outputs.tags }}
           BASE_IMAGE_FULL: ${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
+          MEMPALACE_TOOLKIT_REF: ${{ needs.resolve-versions.outputs.mempalace_toolkit_ref }}
         run: |
           set -euo pipefail
+          BUILD_DATE=$(date -u +%Y-%m-%dT%H:%M:%SZ)
           TAG_FLAGS=()
           while IFS= read -r t; do [[ -n "$t" ]] && TAG_FLAGS+=( -t "$t" ); done <<< "${TAGS}"
           # 3-attempt retry around `docker buildx build --push` (see build-base
@@ -541,7 +430,10 @@ jobs:
                 --build-arg "BASE_IMAGE=${BASE_IMAGE_FULL}" \
                 --build-arg "INSTALL_OPENCODE=true" \
                 --build-arg "INSTALL_OMOS=false" \
-                --build-arg "INSTALL_PI=false" \
+                --build-arg "MEMPALACE_TOOLKIT_REF=${MEMPALACE_TOOLKIT_REF}" \
+                --build-arg "RELEASE_TAG=${RELEASE_TAG}" \
+                --build-arg "BUILD_DATE=${BUILD_DATE}" \
+                --build-arg "SOURCE_REVISION=${GITHUB_SHA:-}" \
                 "${TAG_FLAGS[@]}" \
                 .; then
               echo "==> Attempt ${attempt} succeeded"
@@ -597,8 +489,10 @@ jobs:
           TAGS: ${{ steps.tags.outputs.tags }}
           BASE_IMAGE_FULL: ${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
           OMOS_VERSION: ${{ needs.resolve-versions.outputs.omos_version }}
+          MEMPALACE_TOOLKIT_REF: ${{ needs.resolve-versions.outputs.mempalace_toolkit_ref }}
         run: |
           set -euo pipefail
+          BUILD_DATE=$(date -u +%Y-%m-%dT%H:%M:%SZ)
           TAG_FLAGS=()
           while IFS= read -r t; do [[ -n "$t" ]] && TAG_FLAGS+=( -t "$t" ); done <<< "${TAGS}"
           # 3-attempt retry (see build-base step for rationale). Variant: omos.
@@ -611,238 +505,11 @@ jobs:
                 --build-arg "BASE_IMAGE=${BASE_IMAGE_FULL}" \
                 --build-arg "INSTALL_OPENCODE=true" \
                 --build-arg "INSTALL_OMOS=true" \
-                --build-arg "INSTALL_PI=false" \
                 --build-arg "OMOS_VERSION=${OMOS_VERSION}" \
-                "${TAG_FLAGS[@]}" \
-                .; then
-              echo "==> Attempt ${attempt} succeeded"
-              exit 0
-            fi
-            if [[ "${attempt}" -lt 3 ]]; then
-              backoff=$(( attempt * 15 ))
-              echo "==> Attempt ${attempt} failed, sleeping ${backoff}s before retry"
-              sleep "${backoff}"
-            fi
-          done
-          echo "==> All 3 build+push attempts failed"
-          exit 1
-
-  build-variant-with-pi:
-    needs: [base-decide, smoke-with-pi, resolve-versions]
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - uses: actions/checkout@v4
-      - run: echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-      - run: |
-          rm -rf /opt/hostedtoolcache /opt/microsoft /opt/az /opt/ghc \
-            /usr/local/.ghcup /usr/share/dotnet /usr/share/swift \
-            /usr/local/lib/android /usr/local/share/powershell \
-            /usr/local/share/chromium /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-      - uses: docker/setup-qemu-action@v3
-        with: {platforms: arm64}
-      - uses: docker/setup-buildx-action@v4
-        with: {driver-opts: network=host}
-      - uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-      - name: Compute version-specific tags
-        id: tags
-        run: |
-          VERSION="${{ env.RELEASE_TAG }}"
-          { echo "tags<<EOF"
-            echo "${IMAGE}:${VERSION}-with-pi"
-            if [ "${{ env.PROMOTE_LATEST }}" = "true" ]; then
-              echo "${IMAGE}:latest-with-pi"
-            fi
-            echo "EOF"
-          } >> "$GITHUB_OUTPUT"
-      - name: Build and push variant (with retry)
-        shell: bash
-        env:
-          TAGS: ${{ steps.tags.outputs.tags }}
-          BASE_IMAGE_FULL: ${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
-          PI_VERSION: ${{ needs.resolve-versions.outputs.pi_version }}
-          FORK_REF: ${{ needs.resolve-versions.outputs.fork_ref }}
-          OBSMEM_REF: ${{ needs.resolve-versions.outputs.obsmem_ref }}
-        run: |
-          set -euo pipefail
-          TAG_FLAGS=()
-          while IFS= read -r t; do [[ -n "$t" ]] && TAG_FLAGS+=( -t "$t" ); done <<< "${TAGS}"
-          # 3-attempt retry (see build-base step for rationale). Variant: with-pi.
-          for attempt in 1 2 3; do
-            echo "==> Build+push attempt ${attempt}/3"
-            if docker buildx build \
-                --platform linux/amd64,linux/arm64 \
-                --file Dockerfile.variant \
-                --push \
-                --build-arg "BASE_IMAGE=${BASE_IMAGE_FULL}" \
-                --build-arg "INSTALL_OPENCODE=true" \
-                --build-arg "INSTALL_OMOS=false" \
-                --build-arg "INSTALL_PI=true" \
-                --build-arg "PI_VERSION=${PI_VERSION}" \
-                --build-arg "PI_FORK_REF=${FORK_REF}" \
-                --build-arg "PI_OBSMEM_REF=${OBSMEM_REF}" \
-                "${TAG_FLAGS[@]}" \
-                .; then
-              echo "==> Attempt ${attempt} succeeded"
-              exit 0
-            fi
-            if [[ "${attempt}" -lt 3 ]]; then
-              backoff=$(( attempt * 15 ))
-              echo "==> Attempt ${attempt} failed, sleeping ${backoff}s before retry"
-              sleep "${backoff}"
-            fi
-          done
-          echo "==> All 3 build+push attempts failed"
-          exit 1
-
-  build-variant-omos-with-pi:
-    needs: [base-decide, smoke-omos-with-pi, resolve-versions]
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - uses: actions/checkout@v4
-      - run: echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-      - run: |
-          rm -rf /opt/hostedtoolcache /opt/microsoft /opt/az /opt/ghc \
-            /usr/local/.ghcup /usr/share/dotnet /usr/share/swift \
-            /usr/local/lib/android /usr/local/share/powershell \
-            /usr/local/share/chromium /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-      - uses: docker/setup-qemu-action@v3
-        with: {platforms: arm64}
-      - uses: docker/setup-buildx-action@v4
-        with: {driver-opts: network=host}
-      - uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-      - name: Compute version-specific tags
-        id: tags
-        run: |
-          VERSION="${{ env.RELEASE_TAG }}"
-          { echo "tags<<EOF"
-            echo "${IMAGE}:${VERSION}-omos-with-pi"
-            if [ "${{ env.PROMOTE_LATEST }}" = "true" ]; then
-              echo "${IMAGE}:latest-omos-with-pi"
-            fi
-            echo "EOF"
-          } >> "$GITHUB_OUTPUT"
-      - name: Build and push variant (with retry)
-        shell: bash
-        env:
-          TAGS: ${{ steps.tags.outputs.tags }}
-          BASE_IMAGE_FULL: ${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
-          PI_VERSION: ${{ needs.resolve-versions.outputs.pi_version }}
-          OMOS_VERSION: ${{ needs.resolve-versions.outputs.omos_version }}
-          FORK_REF: ${{ needs.resolve-versions.outputs.fork_ref }}
-          OBSMEM_REF: ${{ needs.resolve-versions.outputs.obsmem_ref }}
-        run: |
-          set -euo pipefail
-          TAG_FLAGS=()
-          while IFS= read -r t; do [[ -n "$t" ]] && TAG_FLAGS+=( -t "$t" ); done <<< "${TAGS}"
-          # 3-attempt retry (see build-base step for rationale). Variant: omos-with-pi.
-          for attempt in 1 2 3; do
-            echo "==> Build+push attempt ${attempt}/3"
-            if docker buildx build \
-                --platform linux/amd64,linux/arm64 \
-                --file Dockerfile.variant \
-                --push \
-                --build-arg "BASE_IMAGE=${BASE_IMAGE_FULL}" \
-                --build-arg "INSTALL_OPENCODE=true" \
-                --build-arg "INSTALL_OMOS=true" \
-                --build-arg "INSTALL_PI=true" \
-                --build-arg "PI_VERSION=${PI_VERSION}" \
-                --build-arg "OMOS_VERSION=${OMOS_VERSION}" \
-                --build-arg "PI_FORK_REF=${FORK_REF}" \
-                --build-arg "PI_OBSMEM_REF=${OBSMEM_REF}" \
-                "${TAG_FLAGS[@]}" \
-                .; then
-              echo "==> Attempt ${attempt} succeeded"
-              exit 0
-            fi
-            if [[ "${attempt}" -lt 3 ]]; then
-              backoff=$(( attempt * 15 ))
-              echo "==> Attempt ${attempt} failed, sleeping ${backoff}s before retry"
-              sleep "${backoff}"
-            fi
-          done
-          echo "==> All 3 build+push attempts failed"
-          exit 1
-
-  build-variant-pi-only:
-    needs: [base-decide, smoke-pi-only, resolve-versions]
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - uses: actions/checkout@v4
-      - run: echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-      - run: |
-          rm -rf /opt/hostedtoolcache /opt/microsoft /opt/az /opt/ghc \
-            /usr/local/.ghcup /usr/share/dotnet /usr/share/swift \
-            /usr/local/lib/android /usr/local/share/powershell \
-            /usr/local/share/chromium /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-      - uses: docker/setup-qemu-action@v3
-        with: {platforms: arm64}
-      - uses: docker/setup-buildx-action@v4
-        with: {driver-opts: network=host}
-      - uses: docker/login-action@v3
-        with:
-          username: ${{ vars.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-      - name: Compute version-specific tags
-        id: tags
-        run: |
-          # Option B: push the pi-only build into the pi-devbox repo as an
-          # internal building-block tag (base-pi-only[-<version>]), NOT under
-          # opencode-devbox. pi-devbox's CI FROMs ${PI_IMAGE}:base-pi-only.
-          VERSION="${{ env.RELEASE_TAG }}"
-          { echo "tags<<EOF"
-            echo "${PI_IMAGE}:base-pi-only-${VERSION}"
-            if [ "${{ env.PROMOTE_LATEST }}" = "true" ]; then
-              echo "${PI_IMAGE}:base-pi-only"
-            fi
-            echo "EOF"
-          } >> "$GITHUB_OUTPUT"
-      - name: Build and push variant (with retry)
-        shell: bash
-        env:
-          TAGS: ${{ steps.tags.outputs.tags }}
-          BASE_IMAGE_FULL: ${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
-          PI_VERSION: ${{ needs.resolve-versions.outputs.pi_version }}
-          FORK_REF: ${{ needs.resolve-versions.outputs.fork_ref }}
-          OBSMEM_REF: ${{ needs.resolve-versions.outputs.obsmem_ref }}
-        run: |
-          set -euo pipefail
-          TAG_FLAGS=()
-          while IFS= read -r t; do [[ -n "$t" ]] && TAG_FLAGS+=( -t "$t" ); done <<< "${TAGS}"
-          # 3-attempt retry (see build-base step for rationale). Variant: pi-only.
-          for attempt in 1 2 3; do
-            echo "==> Build+push attempt ${attempt}/3"
-            if docker buildx build \
-                --platform linux/amd64,linux/arm64 \
-                --file Dockerfile.variant \
-                --push \
-                --build-arg "BASE_IMAGE=${BASE_IMAGE_FULL}" \
-                --build-arg "INSTALL_OPENCODE=false" \
-                --build-arg "INSTALL_OMOS=false" \
-                --build-arg "INSTALL_PI=true" \
-                --build-arg "PI_VERSION=${PI_VERSION}" \
-                --build-arg "PI_FORK_REF=${FORK_REF}" \
-                --build-arg "PI_OBSMEM_REF=${OBSMEM_REF}" \
+                --build-arg "MEMPALACE_TOOLKIT_REF=${MEMPALACE_TOOLKIT_REF}" \
+                --build-arg "RELEASE_TAG=${RELEASE_TAG}" \
+                --build-arg "BUILD_DATE=${BUILD_DATE}" \
+                --build-arg "SOURCE_REVISION=${GITHUB_SHA:-}" \
                 "${TAG_FLAGS[@]}" \
                 .; then
               echo "==> Attempt ${attempt} succeeded"
@@ -863,24 +530,25 @@ jobs:
       - base-decide
       - build-variant-base
       - build-variant-omos
-      - build-variant-with-pi
-      - build-variant-omos-with-pi
-      - build-variant-pi-only
-    # Skip on cache-hit base builds: when need_build=false, base-latest
-    # already points at the same digest as base-<hash>, so the retag is
-    # a tautology and any transient failure of it is purely cosmetic.
-    # Manual workflow_dispatch with promote_latest=true overrides this
-    # gate as an escape hatch (e.g., if base-latest got hand-deleted).
+    # Run on every tag release (and promote_latest=true dispatch). The gate
+    # deliberately does NOT key off need_build anymore: the no-op optimization
+    # for genuine cache-hit releases moved INTO the step as a crane digest
+    # compare (see below). Keying the gate on need_build was wrong because a
+    # prior dry-run dispatch (promote_latest=false) can pre-build+push
+    # base-<hash>, making need_build=false on the subsequent tag run even
+    # though base-latest is still stale — the old gate then skipped promotion
+    # and left base-latest pointing at the PREVIOUS base. (Ported from
+    # pi-devbox b7197e8, which hit exactly this on its v1.2.3 dry-run-first
+    # release, 2026-06-27.)
     #
     # `always()` wrapper + explicit base-variant success check protects
     # against the gitea-Actions default of "skipped need => skip dependent":
-    # a partial-publish run (e.g., omos-with-pi smoke fails) shouldn't
+    # a partial-publish run (e.g., omos smoke fails) shouldn't
     # prevent the base-latest alias from advancing on a real base rebuild.
     if: |
       always() &&
       needs.build-variant-base.result == 'success' &&
-      (inputs.promote_latest == 'true' ||
-       (github.ref_type == 'tag' && needs.base-decide.outputs.need_build == 'true'))
+      (inputs.promote_latest == 'true' || github.ref_type == 'tag')
     runs-on: ubuntu-latest
     container:
       image: catthehacker/ubuntu:act-latest
@@ -905,22 +573,39 @@ jobs:
           crane auth login docker.io \
             -u ${{ vars.DOCKERHUB_USERNAME }} \
             -p "${{ secrets.DOCKERHUB_TOKEN }}"
-      - name: Re-tag base-<hash> as base-latest
+      - name: Re-tag base-<hash> as base-latest (only if stale)
+        env:
+          BASE_HASH_REF: ${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }}
+          BASE_LATEST_REF: ${{ env.IMAGE }}:base-latest
         run: |
-          crane copy \
-            ${{ env.IMAGE }}:${{ needs.base-decide.outputs.base_tag }} \
-            ${{ env.IMAGE }}:base-latest
+          set -euo pipefail
+          # Correctness invariant: after a release, base-latest must resolve to
+          # the SAME digest as the base-<hash> the just-built variants were
+          # FROM. Compare digests rather than trusting need_build — a prior
+          # dry-run dispatch can pre-build base-<hash>, so need_build=false on
+          # the tag run does NOT imply base-latest is already current. When the
+          # digests already match (genuine cache-hit release) this is a no-op,
+          # so we skip the crane copy entirely — preserving the original
+          # "don't do a tautological retag" intent and avoiding any cosmetic
+          # transient-failure exposure on releases that change nothing.
+          want=$(crane digest "${BASE_HASH_REF}")
+          have=$(crane digest "${BASE_LATEST_REF}" 2>/dev/null || echo "")
+          echo "base-<hash> digest: ${want}"
+          echo "base-latest digest: ${have:-<absent>}"
+          if [ "${want}" = "${have}" ]; then
+            echo "base-latest already current; nothing to promote."
+          else
+            echo "Promoting base-latest -> ${BASE_HASH_REF}"
+            crane copy "${BASE_HASH_REF}" "${BASE_LATEST_REF}"
+          fi
 
   # ── Phase 6: update Hub description (only on real release runs) ────
   update-description:
     needs:
       - build-variant-base
       - build-variant-omos
-      - build-variant-with-pi
-      - build-variant-omos-with-pi
-      - build-variant-pi-only
     # Run when at least the base variant published — don't let a single
-    # variant failure (e.g., omos-with-pi smoke threshold) prevent Hub
+    # variant failure (e.g., omos smoke threshold) prevent Hub
     # description refresh for the other variants that did publish.
     # Without this `always()` wrapper, gitea Actions' default behavior
     # of "skipped need => skip dependent" cascades from any failed/
@@ -937,6 +622,24 @@ jobs:
       - uses: actions/checkout@v4
       - name: Update Docker Hub description
         run: |
+          # Substitute {{OPENCODE_VERSION}} placeholders in DOCKER_HUB.md so
+          # the Hub page always shows which opencode version is baked into
+          # :latest. The placeholder lives in DOCKER_HUB.md (committed); CI
+          # fills it at publish time from the pinned ARG in
+          # Dockerfile.variant — the same value that was baked into the
+          # image — so the page and the image never drift. (Mirrors the
+          # {{PI_VERSION}} pattern in pi-devbox's docker-publish.yml.)
+          OPENCODE_VERSION=$(sed -n 's/^ARG OPENCODE_VERSION=//p' Dockerfile.variant | head -1)
+          if [ -z "${OPENCODE_VERSION}" ]; then
+            echo "::error::Could not extract OPENCODE_VERSION from Dockerfile.variant"
+            exit 1
+          fi
+          cp DOCKER_HUB.md /tmp/hub-full.md
+          sed -i "s/{{OPENCODE_VERSION}}/${OPENCODE_VERSION}/g" /tmp/hub-full.md
+          if grep -q '{{OPENCODE_VERSION}}' /tmp/hub-full.md; then
+            echo "::error::DOCKER_HUB.md still contains unsubstituted {{OPENCODE_VERSION}} markers"
+            exit 1
+          fi
           TOKEN=$(curl -s -X POST https://hub.docker.com/v2/auth/token \
             -H "Content-Type: application/json" \
             -d '{"identifier":"${{ vars.DOCKERHUB_USERNAME }}","secret":"${{ secrets.DOCKERHUB_TOKEN }}"}' \
@@ -946,7 +649,7 @@ jobs:
             exit 1
           fi
           HTTP_CODE=$(jq -n \
-            --rawfile full DOCKER_HUB.md \
+            --rawfile full /tmp/hub-full.md \
             --arg short "Portable AI dev environment for opencode. Debian-based with git, Node.js, AWS CLI, and SSH support." \
             '{"full_description": $full, "description": $short}' | \
             curl -s -o /tmp/hub-response.txt -w "%{http_code}" -X PATCH \

.gitea/workflows/lint.yml

@@ -0,0 +1,116 @@
+name: Lint
+
+# The repo's cheap-checks workflow: everything that can be verified WITHOUT
+# building an image. Runs on every push and PR, in ~30s, independent of the
+# build pipeline. Three jobs: workflow lint (actionlint + the Gitea shell
+# guard), Dockerfile lint (hadolint), and the DOCKER_HUB.md docs-sync check.
+#
+# Deliberately the ONLY workflow that runs on a push to main. validate.yml
+# (amd64 build + smoke) is PR/dispatch-only and docker-publish-split.yml is
+# tag-only, so pushing work-in-progress to main never starts an image build.
+#
+# Its original purpose, still the most important one: the "bash-only syntax
+# under the default `sh`/dash shell" footgun. Ported from pi-devbox, where this
+# class broke resolve-versions (ed49b8d) and promote-base-latest (b7197e8 → run
+# 418). actionlint runs shellcheck against each `run:` step using its
+# *effective* shell, so `set -o pipefail` under dash is flagged as SC3040
+# before any expensive build runs.
+on:
+  push:
+  pull_request:
+  workflow_dispatch:
+
+concurrency:
+  group: lint-${{ github.ref }}
+  cancel-in-progress: true
+
+defaults:
+  run:
+    shell: bash
+
+jobs:
+  actionlint:
+    runs-on: ubuntu-latest
+    container:
+      image: catthehacker/ubuntu:act-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install shellcheck
+        run: |
+          apt-get update
+          apt-get install -y --no-install-recommends shellcheck python3-yaml
+
+      - name: Gitea shell guard (catches the actionlint blind spot)
+        # actionlint models GitHub Actions, where the default run shell is
+        # bash, so it does NOT flag bash syntax in a step that merely OMITS
+        # `shell:` — which is exactly how ed49b8d and b7197e8 manifested on
+        # Gitea (default sh/dash). This guard enforces that every run: step
+        # resolves to bash under Gitea's real defaults. Run it BEFORE
+        # actionlint so the more precise diagnostic surfaces first.
+        run: bash scripts/check-workflow-shell.sh .gitea/workflows
+
+      - name: Install actionlint (pinned)
+        env:
+          ACTIONLINT_VERSION: 1.7.7
+        run: |
+          curl -fsSL \
+            "https://github.com/rhysd/actionlint/releases/download/v${ACTIONLINT_VERSION}/actionlint_${ACTIONLINT_VERSION}_linux_amd64.tar.gz" \
+            | tar -xz -C /usr/local/bin actionlint
+          actionlint --version
+
+      - name: Run actionlint
+        # SHELLCHECK_OPTS excludes pure-style codes (quoting/style opinions)
+        # so the guard stays focused on correctness bugs — crucially the
+        # SC3xxx "not POSIX / wrong shell" family that catches the pipefail
+        # footgun. Do NOT exclude SC3040 (set -o pipefail under sh) or any
+        # other SC3xxx code.
+        env:
+          SHELLCHECK_OPTS: "-e SC2086 -e SC2016 -e SC2129 -e SC2001 -e SC2312"
+        # Pass explicit paths: actionlint's no-arg mode auto-detects a
+        # project by looking for `.github/workflows`, which doesn't exist in
+        # this `.gitea/workflows` repo and hard-fails with exit 3
+        # ("no project was found"). Globbing the workflow files is the
+        # supported way to lint a non-GitHub layout.
+        run: actionlint -color .gitea/workflows/*.yml
+
+  hadolint:
+    # Lint the two Dockerfiles that ARE the project (the shell/actions linting
+    # above never looked at them). Config — ignored rules + failure threshold
+    # — lives in .hadolint.yaml, which hadolint reads automatically, so a local
+    # `hadolint Dockerfile.base` reproduces CI exactly.
+    runs-on: ubuntu-latest
+    container:
+      image: catthehacker/ubuntu:act-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Install hadolint (pinned)
+        env:
+          HADOLINT_VERSION: 2.14.0
+        run: |
+          curl -fsSL \
+            "https://github.com/hadolint/hadolint/releases/download/v${HADOLINT_VERSION}/hadolint-Linux-x86_64" \
+            -o /usr/local/bin/hadolint
+          chmod +x /usr/local/bin/hadolint
+          hadolint --version
+
+      - name: Run hadolint
+        run: hadolint Dockerfile.base Dockerfile.variant
+
+  docs-check:
+    # Fails if DOCKER_HUB.md is out of sync with what generate-dockerhub-md.py
+    # would produce from HUB_TEMPLATE. Keeps the two docs from drifting.
+    #
+    # Lives here rather than in validate.yml because it needs no image: keeping
+    # it in the cheap workflow means it still runs on every push to main now
+    # that validate.yml is PR/dispatch-only. Reproduce locally with
+    # `python3 scripts/generate-dockerhub-md.py --check`.
+    runs-on: ubuntu-latest
+    container:
+      image: catthehacker/ubuntu:act-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Check DOCKER_HUB.md is in sync with HUB_TEMPLATE
+        run: python3 scripts/generate-dockerhub-md.py --check

.gitea/workflows/validate.yml

@@ -1,9 +1,24 @@
 name: Validate
 
-# Lightweight validation on pushes to main. Builds single-arch (amd64),
-# runs the smoke test, and checks image size — without pushing anything
-# to Docker Hub. Tag pushes are handled by docker-publish-split.yml which
-# does the full multi-arch split-base build-and-push.
+# Image build + smoke test. This workflow has NO push trigger at all — it fires
+# only on pull_request and on explicit workflow_dispatch.
+#
+# Division of labour across this repo's three workflows:
+#   lint.yml                 — cheap checks, every push/PR: workflow lint,
+#                              Dockerfile lint, DOCKER_HUB.md sync. No builds.
+#   validate.yml (this file) — amd64 build + smoke test. PR / manual only.
+#   docker-publish-split.yml — the release path, tag-only: multi-arch build,
+#                              smoke gates, then push + promote.
+#
+# Net effect: no image build can start from a push to main. A tag — or an
+# explicit dispatch — is required. That is safe because the release path fails
+# closed: variant tags are pushed only after smoke-base/smoke-omos pass, and
+# base-latest is promoted last, so an aborted release leaves at worst an
+# unreferenced base-<hash> blob on Hub, never a half-published version tag.
+#
+# To validate before tagging: open a PR, dispatch this workflow, or dispatch
+# docker-publish-split.yml against a throwaway tag with promote_latest=false
+# (the only route that also exercises a CHANGED BASE — see the trade-off below).
 #
 # Trade-off: variant builds here use the published `base-latest` image
 # from Docker Hub as their parent, NOT a locally-built base. This is
@@ -22,43 +37,35 @@ name: Validate
 # blind-spot applies.
 
 on:
-  push:
-    branches:
-      - main
-    paths-ignore:
-      - 'CHANGELOG.md'
-      - 'README.md'
-      - 'DOCKER_HUB.md'
-      - 'deploy/**'
-      - '.gitleaks.toml'
   pull_request:
     branches:
       - main
+  workflow_dispatch:
+
+# Gitea Actions' default step shell is `sh` (dash); force bash workflow-wide so
+# no run: step silently falls through to dash. Enforced by lint.yml's
+# scripts/check-workflow-shell.sh guard, which scans ALL .gitea/workflows/*.yml
+# (so this file must resolve to bash too, not just docker-publish-split.yml).
+defaults:
+  run:
+    shell: bash
 
 jobs:
-  docs-check:
-    # Fails if DOCKER_HUB.md is out of sync with what generate-dockerhub-md.py
-    # would produce from README.md. Keeps the two docs from drifting.
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Check DOCKER_HUB.md is in sync with README.md
-        run: |
-          python3 scripts/generate-dockerhub-md.py --check
-
   base-change-warning:
     # Surfaces a warning when this commit changes base-image inputs
-    # (Dockerfile.base, rootfs/, entrypoint*.sh). validate.yml uses
-    # Hub's base-latest as the parent for variant builds, so changes to
-    # those files are NOT exercised here — only release tags rebuild the
-    # base via docker-publish-split.yml.
+    # (Dockerfile.base, rootfs/, entrypoint*.sh) AND exports `base_changed` so
+    # validate-base/validate-omos can skip. validate.yml uses Hub's base-latest
+    # as the parent for variant builds, so a changed base is NOT exercised here
+    # — worse, if the same commit tightens smoke-test.sh in lockstep with the
+    # base change (as the nano/micro commit did), those jobs HARD-FAIL against
+    # the stale base-latest: a false red that self-heals only once the release
+    # rebuilds base-latest. So we skip them here and defer full base validation
+    # to the release path (docker-publish-split.yml on tag push).
     runs-on: ubuntu-latest
     container:
       image: catthehacker/ubuntu:act-latest
+    outputs:
+      base_changed: ${{ steps.detect.outputs.base_changed }}
     steps:
       - name: Checkout
         uses: actions/checkout@v4
@@ -66,18 +73,38 @@ jobs:
           fetch-depth: 2
 
       - name: Detect base-input changes
+        id: detect
         run: |
           set -e
-          if ! git diff --name-only HEAD~1 HEAD 2>/dev/null \
-              | grep -qE '^(Dockerfile\.base|rootfs/|entrypoint.*\.sh)$'; then
-            echo "No base-image inputs changed in this commit — validate.yml fully exercises the published base-latest."
-            exit 0
-          fi
-          echo "::warning::This commit changes base-image inputs (Dockerfile.base, rootfs/, or entrypoint*.sh). validate.yml uses Hub's base-latest as the parent for variant builds, so the new base is NOT exercised by this workflow. Cut a release tag, or run a workflow_dispatch of docker-publish-split.yml against a test tag (e.g. v0.0.0-base-test, promote_latest=false) for end-to-end validation of the new base."
+          # Base inputs baked into Dockerfile.base: Dockerfile.base itself, any
+          # file under rootfs/ (COPYed into the base), and entrypoint*.sh. NB:
+          # rootfs/ is a PREFIX match — the earlier '^(...|rootfs/|...)$' anchored
+          # the whole group, so the rootfs/ branch only matched a file literally
+          # named 'rootfs/' and never real paths like rootfs/usr/.../x.py.
+          base_re='^(Dockerfile\.base$|rootfs/|entrypoint.*\.sh$)'
+          changed="$(git diff --name-only HEAD~1 HEAD 2>/dev/null | grep -E "$base_re" || true)"
+          if [ -n "$changed" ]; then
+            echo "base_changed=true" >> "$GITHUB_OUTPUT"
+            echo "::warning::This commit changes base-image inputs (Dockerfile.base, rootfs/, or entrypoint*.sh). validate.yml uses Hub's base-latest as the parent for variant builds, so the new base is NOT exercised by this workflow — validate-base/validate-omos are SKIPPED to avoid a false failure against the stale base-latest. Cut a release tag, or run a workflow_dispatch of docker-publish-split.yml against a test tag (e.g. v0.0.0-base-test, promote_latest=false) for end-to-end validation of the new base."
             echo "Changed base-input files:"
-          git diff --name-only HEAD~1 HEAD | grep -E '^(Dockerfile\.base|rootfs/|entrypoint.*\.sh)$'
+            printf '%s\n' "$changed"
+          else
+            echo "base_changed=false" >> "$GITHUB_OUTPUT"
+            echo "No base-image inputs changed in this commit — validate-base/validate-omos will build against the published base-latest."
+          fi
 
   validate-base:
+    needs: [base-change-warning]
+    # Two gates, both must hold for this job to run:
+    #  1. base_changed != true — the documented blind spot: this workflow builds
+    #     variants on top of Hub's base-latest, so a commit changing base inputs
+    #     would either not exercise the change or hard-fail against a stale
+    #     base-latest (see base-change-warning).
+    #  2. event_name != push — belt-and-braces. Redundant today (this workflow
+    #     has no push trigger), kept deliberately so that re-adding a push
+    #     trigger later cannot silently re-enable image builds on every push to
+    #     main. If you intend that, remove this clause explicitly.
+    if: ${{ needs.base-change-warning.outputs.base_changed != 'true' && github.event_name != 'push' }}
     runs-on: ubuntu-latest
     container:
       image: catthehacker/ubuntu:act-latest
@@ -139,6 +166,9 @@ jobs:
           bash scripts/smoke-test.sh opencode-devbox:ci-base --variant base
 
   validate-omos:
+    needs: [base-change-warning]
+    # Same two gates as validate-base — see the comment there.
+    if: ${{ needs.base-change-warning.outputs.base_changed != 'true' && github.event_name != 'push' }}
     runs-on: ubuntu-latest
     container:
       image: catthehacker/ubuntu:act-latest
@@ -195,179 +225,3 @@ jobs:
       - name: Smoke test
         run: |
           bash scripts/smoke-test.sh opencode-devbox:ci-omos --variant omos
-
-  validate-with-pi:
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Force IPv4 for Docker Hub
-        run: |
-          echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-
-      - name: Reclaim runner disk
-        run: |
-          set -x
-          df -h / || true
-          rm -rf \
-            /opt/hostedtoolcache \
-            /opt/microsoft \
-            /opt/az \
-            /opt/ghc \
-            /usr/local/.ghcup \
-            /usr/share/dotnet \
-            /usr/share/swift \
-            /usr/local/lib/android \
-            /usr/local/share/powershell \
-            /usr/local/share/chromium \
-            /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          apt-get clean || true
-          rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* || true
-          docker system df || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-          df -h / || true
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v4
-        with:
-          driver-opts: network=host
-
-      - name: Build with-pi image (amd64, load to local daemon)
-        uses: docker/build-push-action@v7
-        with:
-          context: .
-          file: Dockerfile.variant
-          platforms: linux/amd64
-          push: false
-          load: true
-          build-args: |
-            BASE_IMAGE=joakimp/opencode-devbox:base-latest
-            INSTALL_PI=true
-          tags: opencode-devbox:ci-with-pi
-
-      - name: Smoke test
-        run: |
-          bash scripts/smoke-test.sh opencode-devbox:ci-with-pi --variant with-pi
-
-  validate-omos-with-pi:
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Force IPv4 for Docker Hub
-        run: |
-          echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-
-      - name: Reclaim runner disk
-        run: |
-          set -x
-          df -h / || true
-          rm -rf \
-            /opt/hostedtoolcache \
-            /opt/microsoft \
-            /opt/az \
-            /opt/ghc \
-            /usr/local/.ghcup \
-            /usr/share/dotnet \
-            /usr/share/swift \
-            /usr/local/lib/android \
-            /usr/local/share/powershell \
-            /usr/local/share/chromium \
-            /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          apt-get clean || true
-          rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* || true
-          docker system df || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-          df -h / || true
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v4
-        with:
-          driver-opts: network=host
-
-      - name: Build omos+with-pi image (amd64, load to local daemon)
-        uses: docker/build-push-action@v7
-        with:
-          context: .
-          file: Dockerfile.variant
-          platforms: linux/amd64
-          push: false
-          load: true
-          build-args: |
-            BASE_IMAGE=joakimp/opencode-devbox:base-latest
-            INSTALL_OMOS=true
-            INSTALL_PI=true
-          tags: opencode-devbox:ci-omos-with-pi
-
-      - name: Smoke test
-        run: |
-          bash scripts/smoke-test.sh opencode-devbox:ci-omos-with-pi --variant omos-with-pi
-
-  validate-pi-only:
-    runs-on: ubuntu-latest
-    container:
-      image: catthehacker/ubuntu:act-latest
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-
-      - name: Force IPv4 for Docker Hub
-        run: |
-          echo 'precedence ::ffff:0:0/96  100' >> /etc/gai.conf
-
-      - name: Reclaim runner disk
-        run: |
-          set -x
-          df -h / || true
-          rm -rf \
-            /opt/hostedtoolcache \
-            /opt/microsoft \
-            /opt/az \
-            /opt/ghc \
-            /usr/local/.ghcup \
-            /usr/share/dotnet \
-            /usr/share/swift \
-            /usr/local/lib/android \
-            /usr/local/share/powershell \
-            /usr/local/share/chromium \
-            /usr/local/share/boost \
-            /usr/lib/jvm 2>/dev/null || true
-          apt-get clean || true
-          rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* || true
-          docker system df || true
-          docker system prune -af --volumes || true
-          docker builder prune -af || true
-          df -h / || true
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v4
-        with:
-          driver-opts: network=host
-
-      - name: Build pi-only image (amd64, load to local daemon)
-        uses: docker/build-push-action@v7
-        with:
-          context: .
-          file: Dockerfile.variant
-          platforms: linux/amd64
-          push: false
-          load: true
-          build-args: |
-            BASE_IMAGE=joakimp/opencode-devbox:base-latest
-            INSTALL_OPENCODE=false
-            INSTALL_PI=true
-          tags: opencode-devbox:ci-pi-only
-
-      - name: Smoke test
-        run: |
-          bash scripts/smoke-test.sh opencode-devbox:ci-pi-only --variant pi-only

.hadolint.yaml

@@ -0,0 +1,27 @@
+# hadolint configuration for opencode-devbox.
+#
+# Both Dockerfiles are linted in CI (.gitea/workflows/lint.yml → `hadolint`
+# job). hadolint reads this file automatically, so a local
+# `hadolint Dockerfile.base` reproduces CI exactly.
+#
+# The ignores below are DELIBERATE project choices — they mirror the
+# philosophy of the shellcheck excludes already applied to `run:` steps
+# (SHELLCHECK_OPTS in lint.yml). Anything NOT listed here still fails the
+# build at `warning` and above, so new Dockerfile smells are caught going
+# forward.
+ignored:
+  - DL3008  # "pin apt versions" — intentionally unpinned: the base tracks
+            #   Debian stable and runs `apt-get upgrade`, so pinning point
+            #   versions would rot and fight security updates.
+  - DL3016  # "pin npm versions" — opencode / oh-my-opencode-slim ARE pinned,
+            #   but via build-args (CI-resolved from npm), not the npm CLI.
+  - DL4006  # "set -o pipefail before a pipe" — the piped RUNs are
+            #   download|extract steps with their own retries / `set -e`.
+            #   Switching the global SHELL to bash is a larger, base-affecting
+            #   change — tracked in IDEAS.md.
+  - DL3003  # "use WORKDIR, not cd" — cosmetic in the few `cd` RUNs here.
+  - SC2086  # "double-quote to prevent word-splitting" — the same code is
+            #   excluded for shell `run:` steps in lint.yml; splitting is
+            #   intentional in these contexts.
+
+failure-threshold: warning

AGENTS.md

@@ -4,39 +4,66 @@
 
 Docker image packaging [opencode](https://opencode.ai) into a production-ready dev container. Image variants are published to Docker Hub via Gitea Actions CI. Not a library or application — this is infrastructure (Dockerfiles, entrypoint scripts, docker-compose, documentation).
 
+> **pi was removed in v2.0.0** (deprecated since v1.17.2). The `INSTALL_PI`
+> build arg, the `with-pi` / `omos-with-pi` / `pi-only` variants, the
+> `base-pi-only` published tag, and all `~/.pi`-related wiring are gone. pi
+> now ships from its own repo (`joakimp/pi-devbox`). Do not add pi
+> functionality here. The removal history + the `NPM_CONFIG_PREFIX`
+> relocation (`~/.pi/npm-global` → `~/.config/opencode/npm-global`, with a
+> one-time migration shim in `entrypoint-user.sh`) are recorded in
+> `docs/CLEANUP-v2.0.0.md` and the v2.0.0 CHANGELOG entry.
+
 ## File roles
 
 - `Dockerfile.base` — variant-independent layers (apt, locales, AWS CLI, Node.js, mempalace, gitea-mcp, user setup, chromadb prewarm, ENVs, entrypoints). Published as `joakimp/opencode-devbox:base-<sha12>`. Rebuilt only when its content hash changes.
-- `Dockerfile.variant` — `FROM`s the base and adds only opencode/omos/pi installs gated by build args: `INSTALL_OPENCODE` (default true), `INSTALL_OMOS`, `INSTALL_PI`, and `INSTALL_MEMPALACE`. All GitHub-sourced binaries are pinned with version ARGs. When `INSTALL_PI=true` it also clones `pi-fork` + `pi-observational-memory` (from `github.com/elpapi42`, refs `PI_FORK_REF`/`PI_OBSMEM_REF`) to `/opt` and runs `npm install` there at build time so the `fork`/`recall` extensions can load (a local-path `pi install` does not npm-install). The `pi-only` variant sets `INSTALL_OPENCODE=false`, `INSTALL_PI=true` — pi without opencode, the single source of truth for the separate `pi-devbox` image. It is built and smoke-tested here, but **published into the `joakimp/pi-devbox` repo** as the internal building-block tag `base-pi-only[-vX.Y.Z]` (NOT under `opencode-devbox`), so an opencode-devbox tag never ships without opencode.
-- `entrypoint.sh` — runs as root: UID/GID adjustment, SSH permissions, volume ownership fixes (skipped via `.devbox-owner` sentinel when ownership is already correct). Then drops to developer via gosu. Volume ownership loop covers `~/.pi/` when `INSTALL_PI=true`.
-- `entrypoint-user.sh` — runs as developer: git config, opencode.jsonc generation (delegated to `generate-config.py`), LAN-access setup (delegated to `setup-lan-access.sh`), pi-toolkit + pi-extensions deploy (when pi installed), pi settings.json bootstrap, mempalace pi-bridge symlink, runtime `pi install /opt/{pi-fork,pi-observational-memory}` registration (idempotent), skillset auto-deploy from mounted skillset repo, OMOS setup.
-- `rootfs/usr/local/lib/opencode-devbox/setup-lan-access.sh` — host-OS-agnostic LAN reachability helper. Detects VM-backed hosts (macOS OrbStack / Docker Desktop, via `host.docker.internal` resolution) and generates a writable `~/.ssh-local/config` using the host as an SSH jump; no-op on native Linux. Controlled by `DEVBOX_LAN_ACCESS` / `HOST_SSH_USER` / `DEVBOX_HOST_ALIAS`. Ships the mechanism only (generic `host` jump alias); user targets stay in their bind-mounted `~/.ssh/config`. Non-fatal. Counted in the base hash, so editing it advances `base-latest`.
+- `Dockerfile.variant` — `FROM`s the base and adds only opencode/omos installs gated by build args: `INSTALL_OPENCODE` (default true), `INSTALL_OMOS`, and `INSTALL_MEMPALACE`. All GitHub-sourced binaries are pinned with version ARGs. Two variants: `base` (`INSTALL_OPENCODE=true`) and `omos` (`+INSTALL_OMOS=true`).
+- `entrypoint.sh` — runs as root: UID/GID adjustment, SSH permissions, volume ownership fixes (skipped via `.devbox-owner` sentinel when ownership is already correct). Then drops to developer via gosu.
+- `entrypoint-user.sh` — runs as developer: git config, opencode.jsonc generation (delegated to `generate-config.py`), LAN-access setup (delegated to `setup-lan-access.sh`), a one-time npm-global prefix migration shim (legacy `~/.pi/npm-global` → `~/.config/opencode/npm-global`), skillset auto-deploy from mounted skillset repo, OMOS bundled-skills reconcile (symlinks the image's bundled skills into `~/.agents/skills/`), image-baked fallback-skills reconcile (symlinks `/usr/local/share/opencode-devbox/skills/*` into `~/.agents/skills/` only-when-absent) + harness-instruction reconcile (symlinks `/usr/local/share/opencode-devbox/instructions/*.md` into `~/.config/opencode/instructions/`), OMOS config setup.
+- `rootfs/usr/local/lib/opencode-devbox/setup-lan-access.sh` — host-OS-agnostic LAN reachability helper. Always writes the writable `~/.ssh-local/config` sidecar on **every** host OS: a `Host *` block that redirects `ControlPath` into `~/.ssh-local/cm/` (first-value-wins over any read-only `~/.ssh`-bound per-host setting) plus `Include ~/.ssh/config`. On VM-backed hosts (macOS OrbStack / Docker Desktop, detected via `host.docker.internal` resolution) it additionally inserts the host-jump block; on native Linux that block is omitted (LAN is reachable directly) but the sidecar is still rendered. Previously the script exited early on native Linux, leaving `dssh`/`dscp` broken when `~/.ssh` was read-only there. Controlled by `DEVBOX_LAN_ACCESS` / `HOST_SSH_USER` / `DEVBOX_HOST_ALIAS` / `DEVBOX_LAN_AUTOJUMP_PRIVATE`. Ships the mechanism only (generic `host` jump alias); user targets stay host-side — named-peer `ProxyJump host` overrides go in a bind-mounted `~/.config/devbox-shell/ssh-lan.conf` (Included before `~/.ssh/config`), never baked into the image. **Scoping invariant:** every `Include` in the generated config MUST be preceded by a bare `Host *` reset — an `Include` is scoped to the enclosing `Host`/`Match` block, so without the reset the included config only applies when targeting `host`/`mac` and named peers fall back to SSH defaults. Non-fatal. Counted in the base hash, so editing it advances `base-latest`.
 - `rootfs/usr/local/lib/opencode-devbox/generate-config.py` — generates `~/.config/opencode/opencode.jsonc` from env vars. Never overwrites an existing config (checks both `.json` and `.jsonc`). Auto-registers MCP servers for detected tools (mempalace via `mempalace-mcp`, gitea-mcp, context7 remote endpoint).
+- `rootfs/usr/local/bin/opencode-devbox-version` — reader for the build manifest that `Dockerfile.variant` bakes at `/etc/opencode-devbox/build-manifest.json`. Three modes (`--json`, `--quiet`, default human) plus a live-vs-baked `opencode --version` drift check (a user `npm install -g opencode` lands on the persistent config volume and can shadow the baked `/usr` binary). Printed as the first line of `entrypoint-user.sh` so "which image am I in?" is answered at start. Added v2.9.0 — before that the manifest was baked but nothing read it. Lives under `rootfs/`, so editing it advances the base content hash.
 - `scripts/smoke-test.sh` — post-build image verification. Asserts binary presence, opencode startup, entrypoint correctness, config generation idempotency, and image size thresholds. Used by both CI workflows.
+- `scripts/recreate-sanity-check.sh` — **runtime** post-recreate verification (counterpart to the build-time `smoke-test.sh`). Run inside the container after `docker compose up -d --force-recreate` to confirm the new image is live (opencode version matches `Dockerfile.variant`'s `OPENCODE_VERSION`), persisted named volumes survived (mempalace palace, opencode.db, bash-history), omos runtime skill symlinks resolve, shell defaults re-seeded, and `/opt` toolkits intact. Not run by CI or the entrypoint — it needs the running container + volumes that smoke-test.sh (which uses `--entrypoint=""`) cannot see.
 - `scripts/generate-dockerhub-md.py` — generates `DOCKER_HUB.md` from a hand-maintained `HUB_TEMPLATE` constant. `--check` fails if the committed file is out of sync (enforced by the `validate` workflow).
 - `DOCKER_HUB.md` — **auto-generated** from `HUB_TEMPLATE` in `scripts/generate-dockerhub-md.py`. Do not edit directly. Pushed to Docker Hub description via CI API call. Must stay under 25 kB. Short description field must be ≤100 bytes.
 - `README.md` — authoritative source documentation for everything in this repo. Independent of `DOCKER_HUB.md`: the Hub doc is hand-maintained in the generator's `HUB_TEMPLATE` and intentionally slim, linking back to the gitea README for depth.
 - `.gitea/README.md` — **read this first** if you're touching CI. Architectural overview of the build pipeline (production vs split-base), wall-clock estimates, NPM_CONFIG_PREFIX gotcha, runner expectations, migration plan.
-- `.gitea/workflows/validate.yml` — lightweight amd64 build + smoke test on push to main and PRs. Also runs the DOCKER_HUB.md sync check.
-- `.gitea/workflows/docker-publish-split.yml` — production CI pipeline on tag push (`v*`). Two-phase split-base: computes base hash, conditionally builds base, runs 5 parallel smoke tests, then 5 parallel multi-arch variant builds, promotes `base-latest` alias, updates Docker Hub description.
+- `.gitea/workflows/validate.yml` — amd64 build + smoke test for both variants. **No push trigger at all**: `pull_request` and `workflow_dispatch` only, so it never runs on a push to main. See the trigger-model convention below.
+- `.gitea/workflows/docker-publish-split.yml` — production CI pipeline on tag push (`v*`). Two-phase split-base: computes base hash, conditionally builds base, runs 2 parallel smoke tests, then 2 parallel multi-arch variant builds, promotes `base-latest` alias, updates Docker Hub description.
+- `.gitea/workflows/lint.yml` — the repo's cheap-checks workflow (~30s) and the **only** workflow that runs on a push to main. Three jobs: (1) `actionlint` — a Gitea-accurate shell guard (`scripts/check-workflow-shell.sh`) plus pinned actionlint + shellcheck, asserting every `run:` step resolves to `bash` under Gitea's `sh` default and closing the actionlint blind spot (actionlint models GitHub, whose default `run` shell is bash); (2) `hadolint` — pinned 2.14.0 against both Dockerfiles, config in `.hadolint.yaml` so a local run reproduces CI; (3) `docs-check` — `DOCKER_HUB.md` vs `HUB_TEMPLATE` sync, moved here from `validate.yml` so it survives that workflow becoming PR/dispatch-only. Ported from pi-devbox.
 
 ## Versioning scheme
 
-Tags follow `v{opencode_version}[letter]` — e.g. `v1.14.20` for the first build on a new opencode release, and `v1.14.20b`, `v1.14.20c`, … for subsequent rebuilds on the same opencode version.
+Image tags follow **independent semver** — they version *this image*, not the
+bundled opencode release. **`v2.0.0` is the decoupling point** (the pi-removal
+breaking release); from there the opencode npm version is tracked in
+`CHANGELOG.md` and the `OPENCODE_VERSION` ARG but no longer drives the tag. This
+mirrors the sibling [`joakimp/pi-devbox`](https://gitea.jordbo.se/joakimp/pi-devbox)
+repo, which decoupled from the pi tool version at its own `v1.0.0`.
 
-- The number tracks the opencode npm version (see `OPENCODE_VERSION` ARG in `Dockerfile.variant`).
-- **No letter suffix** on the first build of a new opencode version — the bare `v{opencode_version}` tag is the canonical release.
-- **Letter suffix is the build ordinal**, starting at `b` for the second build. The letter `a` is **never used** — think of the suffix as counting rebuilds: `b = 2nd, c = 3rd, d = 4th, …`. For opencode version `1.14.20`: first build `v1.14.20`, second `v1.14.20b`, third `v1.14.20c`, and so on.
-- A letter suffix is only used for container-level rebuilds — tooling changes, CVE fixes, doc-driven rebuilds, entrypoint bugfixes — that don't change the underlying opencode version.
-- **Pre-flight check before cutting any non-letter-suffixed tag** — verify the bump is real:
+- **MAJOR** — breaking changes to how users run/configure the container (volume
+  layout, removed variants/build-args, an entrypoint contract change that
+  requires user action). `v2.0.0` (pi removal + npm-prefix relocation) is the
+  reference example.
+- **MINOR** — backward-compatible features: new variants/tags, new opt-in
+  behavior, new env vars, or changed-but-compatible semantics. Example: `v2.1.0`
+  added the OMOS bundled-skills image-symlink mechanism; `v2.3.0` added the
+  image-baked fallback skills + harness-instruction mechanism.
+- **PATCH** — opencode/tool version bumps and small fixes that don't change the
+  contract. When a release pairs a tool bump with a feature, the feature wins
+  and it's a minor.
+- **Pre-flight check** — **before triggering any release CI build, check whether opencode is behind and offer to bump it.** Compare the pinned `OPENCODE_VERSION` in `Dockerfile.variant` against the latest npm release (`./check-versions.sh Dockerfile.variant` flags `→ X available` for opencode — note the base-only pins like `GOSU_VERSION` live in `Dockerfile.base`, so run it against that too for a full sweep). **If the latest stable `opencode-ai` is newer than the pin, ASK THE USER whether to bump before the build — never bump silently;** a version bump is a functional change they may want to hold or schedule. When a bump *is* part of the release, re-check the `OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS` removal-trigger against the new version's release notes (see the bg-subagents note under *Critical conventions*), then verify the pin is real before claiming it in the CHANGELOG:
   ```bash
-  npm view opencode-ai version           # must equal the X.Y.Z in your tag
+  npm view opencode-ai version           # must equal the X.Y.Z you pin in Dockerfile.variant
   ```
-  If the npm version equals the *previous* release's `X.Y.Z`, you're cutting a letter-suffix rebuild (`vX.Y.Zc`, `vX.Y.Zd`, …), not a new minor. **A bare `vX.Y.Z` tag is a claim that opencode upstream just released `X.Y.Z`** — if that claim is wrong, future opencode releases will collide with your tag namespace and the version-tracking story breaks.
+  Historical note: under the *old* `v{opencode_version}[letter]` scheme a
+  mismatched tag was a namespace hazard — e.g. `v1.15.12` was cut while
+  opencode was still `1.15.11`, then re-cut as `v1.15.11c` (2026-05-28), costing
+  a CI cycle. Semver tags no longer encode the opencode version, so that
+  specific collision class is gone — but a CHANGELOG that names the wrong
+  upstream version is still wrong.
 
-  Cautionary example: 2026-05-28 morning, `v1.15.12` was cut while opencode-ai was still at `1.15.11`. The commit message itself acknowledged "OPENCODE_VERSION stays at 1.15.11" but tagged `v1.15.12` anyway. Re-cut as `v1.15.11c` the same afternoon (see CHANGELOG). The `v1.15.12` git tag and Hub images stayed as historical artifacts; the slip cost a CI cycle and a CHANGELOG-rewrite. **Run the npm view check at the top of every release-day cut.**
-
-CI produces eight Docker Hub tags **under `opencode-devbox`** per release: `vX.Y.Z[n]`, `latest`, `vX.Y.Z[n]-omos`, `latest-omos`, `vX.Y.Z[n]-with-pi`, `latest-with-pi`, `vX.Y.Z[n]-omos-with-pi`, `latest-omos-with-pi` — one tag pair (versioned + floating alias) per opencode-bearing variant (four variants). A fifth build, `pi-only`, is built+smoked here but pushed into the **`joakimp/pi-devbox`** repo as `base-pi-only-vX.Y.Z` (+ `base-pi-only` on tag builds), where it becomes the base for that image.
+CI produces four Docker Hub tags **under `opencode-devbox`** per release: `vX.Y.Z`, `latest`, `vX.Y.Z-omos`, `latest-omos` — one tag pair (versioned + floating alias) per variant (two variants: `base`, `omos`).
 
 When bumping the opencode version, bump `OPENCODE_VERSION` in `Dockerfile.variant` and update the comment in `.env.example` if it names a specific model/version for context.
 
@@ -47,10 +74,9 @@ When drafting a release CHANGELOG entry, pull notes from the **canonical upstrea
 | Package | Canonical upstream | What you'll find there |
 |---|---|---|
 | `opencode-ai` (npm) | <https://github.com/anomalyco/opencode/releases> | Per-version release notes with Core / TUI / Desktop / SDK sections, contributor attributions. Some versions have empty bodies (internal/no-user-visible); most do not. |
-| `@earendil-works/pi-coding-agent` (npm) | The `CHANGELOG.md` shipped inside the npm tarball: `npm pack @earendil-works/pi-coding-agent@<version>` then extract `package/CHANGELOG.md`. | Rich changelog with New Features / Added / Changed / Fixed sections per version. |
 | Other floated tools (gosu, fzf, bat, eza, zoxide, uv, nvim, gitea-mcp, Go, oh-my-opencode-slim) | Each project's own GitHub releases page | Usually less material per release; quote selectively. |
 
-**Trap to avoid:** there is a `github.com/sst/opencode` repo that some search results surface; that's a fork (and probably the historical name people associate with opencode given the upstream lineage). It does NOT track the same release timeline. Use `anomalyco/opencode` for opencode release notes.
+**Where opencode actually lives (read this before you go looking):** the canonical opencode repo is **`github.com/anomalyco/opencode`**. It used to be `github.com/sst/opencode` and was **renamed/moved to `anomalyco/opencode` months ago** — `sst/opencode` is the *same* repo and now issues a `301 → anomalyco/opencode` (verified 2026-06-16). It is **not** a separate fork. Old `sst/opencode` links still resolve via the redirect, but always treat `anomalyco/opencode` as the source of truth for releases, PRs, and issues so search results pointing at the old name don't surprise you. Quick check: `curl -sI https://github.com/sst/opencode | grep -i location` → `anomalyco/opencode`.
 
 Fetch pattern (saved here for muscle memory):
 
@@ -60,9 +86,6 @@ npm view opencode-ai time --json | python3 -c 'import sys,json,re; d=json.load(s
 
 # Release notes for a specific version
 curl -s https://api.github.com/repos/anomalyco/opencode/releases/tags/v1.15.10 | python3 -c 'import sys,json; print(json.load(sys.stdin).get("body","(empty)"))'
-
-# pi changelog
-cd /tmp && npm pack @earendil-works/pi-coding-agent@0.75.5 && tar -xzf earendil-works-pi-coding-agent-0.75.5.tgz package/CHANGELOG.md && head -40 package/CHANGELOG.md
 ```
 
 ## Critical conventions
@@ -78,33 +101,89 @@ cd /tmp && npm pack @earendil-works/pi-coding-agent@0.75.5 && tar -xzf earendil-
   Release-day checklist: README → (regenerate DOCKER_HUB.md only if HUB_TEMPLATE changed) → promote CHANGELOG Unreleased → grep AGENTS.md for stale counts → commit → tag → push tag.
 
   **Between releases the same coupling applies.** Doc drift is not just a release-day concern — a workflow tweak, entrypoint change, or `generate-config.py` refactor can leave any of these four files lying. Before committing a non-release change, grep the docs for references to what you touched: `git diff --name-only HEAD | xargs -I{} grep -l 'thing-you-changed' README.md AGENTS.md DOCKER_HUB.md .gitea/README.md .env.example`. If a doc says "four variants" / "two phases" / "runs on amd64 only" and your change made that no longer true, fix it in the same commit.
-- **GitHub/Gitea-sourced binaries float by default** — gosu, fzf, git-lfs, gitleaks, nvim, bat, eza, zoxide, uv, gitea-mcp, Go, oh-my-opencode-slim all default to `latest`. Each build-time install step reads the `/releases/latest` Location redirect (or the go.dev JSON feed for Go) and derives the concrete version. Use the same `ARCH` case-switch pattern for multi-arch support (amd64/arm64) — mind project-specific arch-name deviations (gitleaks uses `x64`, bat/eza/zoxide use `x86_64`/`aarch64`, gosu uses `amd64`/`arm64`). Intentional pins: `OPENCODE_VERSION` (drives the image tag), `NODE_VERSION=22` (major pin), `DEBIAN_VERSION=trixie-slim` (OS base). Adding a new upstream tool: follow the existing floated-version pattern, don't hardcode a specific tag.
+- **GitHub/Gitea-sourced binaries float by default** — gosu, fzf, git-lfs, gitleaks, nvim, bat, eza, zoxide, uv, gitea-mcp, Go, oh-my-opencode-slim all default to `latest`. Each build-time install step reads the `/releases/latest` Location redirect (or the go.dev JSON feed for Go) and derives the concrete version. Use the same `ARCH` case-switch pattern for multi-arch support (amd64/arm64) — mind project-specific arch-name deviations (gitleaks uses `x64`, bat/eza/zoxide use `x86_64`/`aarch64`, gosu uses `amd64`/`arm64`). Intentional pins: `OPENCODE_VERSION` (drives the image tag), `NODE_VERSION=22` (major pin), `DEBIAN_VERSION=trixie-slim` (OS base). Adding a new upstream tool: follow the existing floated-version pattern, don't hardcode a specific tag — **and add its `*_VERSION` ARG to the README *Build Args* table** (the floated-`latest` row alongside `GOSU_VERSION, FZF_VERSION, …`). That table is a *separate* exhaustive list from the env-var table and has silently drifted before: `MICRO_VERSION`, `TEALDEER_VERSION`, and `TYPST_VERSION` were missing from it for three releases (v2.4.0–v2.7.0), caught only by the 2026-07-13 doc-drift audit.
 - **Resolved versions are logged by the smoke test** — `scripts/smoke-test.sh` prints a "Resolved component versions" table as its first step. CI logs always capture what got baked into a given image even when ARGs default to `latest`.
-- **`PI_VERSION` and `OMOS_VERSION` MUST be passed by CI as concrete versions**, not left at the `latest` default. The npm install steps in `Dockerfile.variant` (`npm install -g @earendil-works/pi-coding-agent` / `oh-my-opencode-slim@${OMOS_VERSION}`) produce identical layer-hashes when the ARG values are byte-identical across builds; combined with the registry buildcache (`base-buildcache`) the layer gets reused even when `latest` would have resolved to a newer upstream. This is the same class of bug that bit pi-devbox v0.74.0 → v0.75.5 (silent same-bytes-across-releases regression discovered 2026-05-23, fixed in pi-devbox v0.75.5b). It is currently *masked* in opencode-devbox by `OPENCODE_VERSION` being a hard-coded ARG that bumps every release — that bump invalidates the parent-chain cache key for the downstream pi/omos layers — but the masking would fail the moment a `vN.N.Nb` opencode-version-unchanged release ships that only bumps pi or omos. Preventative fix: `.gitea/workflows/docker-publish-split.yml` has a `resolve-versions` job that runs `npm view @earendil-works/pi-coding-agent version` and `npm view oh-my-opencode-slim version`, exposing concrete values as outputs that every variant smoke + build job consumes via build-args. Smoke tests assert via `EXPECTED_PI_VERSION` / `EXPECTED_OMOS_VERSION` env vars — would catch the regression on the next release rather than four releases later. **If you change the variant build-args list, the resolve-versions job, or the smoke EXPECTED_*_VERSION wiring, audit all affected jobs in lockstep.**
+- **`OMOS_VERSION` MUST be passed by CI as a concrete version**, not left at the `latest` default. The npm install step in `Dockerfile.variant` (`oh-my-opencode-slim@${OMOS_VERSION}`) produces an identical layer-hash when the ARG value is byte-identical across builds; combined with the registry buildcache (`base-buildcache`) the layer gets reused even when `latest` would have resolved to a newer upstream. This is the same class of bug that bit pi-devbox v0.74.0 → v0.75.5 (silent same-bytes-across-releases regression discovered 2026-05-23, fixed in pi-devbox v0.75.5b). It is currently *masked* in opencode-devbox by `OPENCODE_VERSION` being a hard-coded ARG that bumps every release — that bump invalidates the parent-chain cache key for the downstream omos layer — but the masking would fail the moment a `vN.N.Nb` opencode-version-unchanged release ships that only bumps omos. Preventative fix: `.gitea/workflows/docker-publish-split.yml` has a `resolve-versions` job that runs `npm view oh-my-opencode-slim version`, exposing the concrete value as an output that the omos smoke + build jobs consume via build-args. Smoke tests assert via the `EXPECTED_OMOS_VERSION` env var — would catch the regression on the next release rather than several releases later. **If you change the variant build-args list, the resolve-versions job, or the smoke EXPECTED_*_VERSION wiring, audit all affected jobs in lockstep.**
+- **`resolve-versions` also pins `mempalace-toolkit` to a SHA** — `resolve-versions` resolves the `mempalace-toolkit` `main` HEAD to a commit SHA (`mempalace_toolkit_ref` output) via the gitea commits API (`/api/v1/repos/joakimp/mempalace-toolkit/commits?limit=1&sha=main`; gitea allows **unauthenticated** public-repo listing, so no secret is required). Unlike every other dependency, `mempalace-toolkit` is cloned in **`Dockerfile.base`**, not the variant — so the resolve→build-arg→variant plumbing bypasses it. To make a moved toolkit actually land, the resolved SHA is **folded into the `base-decide` hash** (so `base_tag` changes → base rebuilds) AND passed to `build-base` as `--build-arg MEMPALACE_TOOLKIT_REF`. Consequently **`base-decide` now depends on `resolve-versions`** (they no longer run in parallel), and the base clone uses a SHA-capable `git fetch <ref> + checkout FETCH_HEAD` (a `git clone --branch <40-char-SHA>` would fail). Trade-off: `base_tag` now reflects a live gitea lookup — on an API blip it falls back to `main`, triggering one *extra* base rebuild, never a *missed* one. If you touch `resolve-versions`, `base-decide`'s hash inputs, or the `build-base` build-args, audit `.gitea/README.md` Step 1 in lockstep.
 - **Registry buildkit cache-export is currently disabled** — do NOT re-add `cache-from`/`cache-to` to the `build-base` step in `.gitea/workflows/docker-publish-split.yml` without first verifying that buildkit's `mode=max` cache-export to `registry-1.docker.io` no longer returns HTTP 400 from the Hub CDN edge. The regression surfaced ~2026-05-23 and broke five consecutive opencode-devbox publish attempts (runs #332/333/334/336 + a rerun); root-caused on 2026-05-28 by a manual host-side publish that reproduced the same 400 only on `--cache-to` while image push worked fine. Failure shape is stable (`Offset:0` in the `_state` token, HTML response body = CDN-tier rejection, not registry backend), repo-specific (we're the only repo writing `:base-buildcache` mode=max), and explains why pinning `setup-buildx-action@v4.0.0` didn't help (action pin doesn't change the bundled buildkit version on the catthehacker runner image). Trade-off: dockerfile.base changes pay a full ~3 min rebuild instead of pulling cached layers; unchanged bases short-circuit at the Hub-probe step in `base-decide` and never re-build anyway. Variants don't use registry cache so they're unaffected. Re-enable condition: upstream moby/buildkit fix lands AND a low-risk test run succeeds without 400s. See CHANGELOG v1.15.12 `Unreleased` block for the full diagnostic chain. Manual escape-hatch publish procedure: `docs/manual-host-publish.md`.
-- **Push steps wrap `docker buildx build --push` in a 3-attempt retry loop** (15s, 30s backoff) for transient `registry-1.docker.io` blips — rate limits, brief 5xx, CDN flap. Implemented as inline `shell: bash` steps with `docker buildx build` raw rather than `docker/build-push-action@v7` so the loop is visible and tweakable. Affects the 1 base + 5 variant push steps in `.gitea/workflows/docker-publish-split.yml`; smoke-test builds (`load: true`, no push) are untouched. **This does NOT mask deterministic failures** — a true regression (like the cache-export 400 of 2026-05-23..28) fails all 3 attempts identically and the job still fails. Orthogonal to the cache-export disablement above: cache-export was about a deterministic protocol mismatch, retry is about absorbing genuine transients. Both are belt-and-braces with the `ci-release-watcher` skill's transient-rerun heuristic. If you change the matrix of push steps, keep the retry wrapper consistent across them — the pattern is duplicated rather than factored out because Gitea Actions doesn't support reusable composite shell steps cleanly.
+- **Push steps wrap `docker buildx build --push` in a 3-attempt retry loop** (15s, 30s backoff) for transient `registry-1.docker.io` blips — rate limits, brief 5xx, CDN flap. Implemented as inline `shell: bash` steps with `docker buildx build` raw rather than `docker/build-push-action@v7` so the loop is visible and tweakable. Affects the 1 base + 2 variant push steps in `.gitea/workflows/docker-publish-split.yml`; smoke-test builds (`load: true`, no push) are untouched. **This does NOT mask deterministic failures** — a true regression (like the cache-export 400 of 2026-05-23..28) fails all 3 attempts identically and the job still fails. Orthogonal to the cache-export disablement above: cache-export was about a deterministic protocol mismatch, retry is about absorbing genuine transients. Both are belt-and-braces with the `ci-release-watcher` skill's transient-rerun heuristic. If you change the matrix of push steps, keep the retry wrapper consistent across them — the pattern is duplicated rather than factored out because Gitea Actions doesn't support reusable composite shell steps cleanly.
 - **Shell scripts use `set -euo pipefail`** — both entrypoints are strict. Errors in volume chown or SSH permission operations are intentionally suppressed with `|| true`.
+- **Background subagents flag baked ON — experimental, watch for promotion** — `Dockerfile.base` sets `ENV OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=true`. opencode gates native background subagents behind this flag (`packages/opencode/src/tool/task.ts` fails with `Background subagents require OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=true` when unset); `oh-my-opencode-slim` V2+ makes background orchestration its **default** workflow, so the omos variant is effectively degraded without it. It's a base ENV (applies to both variants; harmless on plain opencode — only *enables* a capability) and stays runtime-overridable (`-e …=false`). It's counted in the base hash, so editing that line advances `base-latest`. **REMOVAL TRIGGER:** when opencode promotes background subagents out of `EXPERIMENTAL_` (flag renamed or made default), drop the ENV. No upstream roadmap date as of
+  opencode 1.18.13 / omos 2.2.9 (2026-08) — re-verified at all three of the
+  1.17.20, 1.18.12 and 1.18.13 tags that `packages/opencode/src/effect/runtime-flags.ts`
+  is unchanged and still gates the flag behind `OPENCODE_EXPERIMENTAL`. Documented in lockstep in README env table, `.env.example`, and asserted by `scripts/smoke-test.sh` (`bg-subagents env baked`).
+- **agent-browser + Chromium is the base's size driver — thresholds are coupled** — `Dockerfile.base` installs the `agent-browser` CLI and a Playwright-managed Chromium (~625 MB after dropping the redundant `chromium_headless_shell-*` build). It is by far the largest single thing in the image and ships in **both** variants. The Chrome binary is reached via the stable symlink `/usr/local/bin/agent-chrome` (exposed as `AGENT_BROWSER_EXECUTABLE_PATH`) and the Dockerfile **`find`s** it rather than hardcoding a path, because Playwright's browser dir is per-version *and* per-arch (`chrome-linux` on arm64, `chrome-linux64` on amd64). If you add anything else large, or remove this layer in a fork, update the size thresholds in `scripts/smoke-test.sh` **in the same commit** — a threshold trip mid-release causes a partial publish and a letter-suffix recovery cycle. Do **not** copy pi-devbox's threshold number: it sums `docker history` while this repo uses `docker image inspect .Size`.
+- **`yq` here means mikefarah's Go v4, not Debian's `yq`** — Debian/Ubuntu's `yq` apt package is the unrelated Python kislyuk/yq (a jq-syntax wrapper on a 3.x line). Since v2.9.0 the apt package is **removed** and the mikefarah binary is installed from GitHub, because the `cloud-init` repo's `provision.sh`/`deploy.sh` require v4 syntax — and because `THIRD_PARTY.md` had credited "yq (mikefarah)" while the image actually shipped the Python one. `scripts/smoke-test.sh` pins the contract with `yq --version | grep -qE 'mikefarah.*version v4'`, so both a regression to the apt package and a future yq v5 fail CI. Do not "simplify" this back into the apt list, and do not install both — with two `yq` binaries on PATH the meaning of `yq` silently depends on PATH order.
+- **MemPalace pin — schema-regression watch target (workaround already removed)** — `MEMPALACE_VERSION` is deliberately pinned (currently **3.6.0**) rather than floated, because an unpinned `uv tool install mempalace` is what once silently swept in a broken `diary_write` schema. History: mempalace 3.3.x/3.4.0 advertised a root-level `anyOf` on `mempalace_diary_write`'s `inputSchema`, which Anthropic's tools API (and Codex) reject at session start (`input_schema does not support oneOf, allOf, or anyOf at the top level`), making the *whole* MCP server fail to load. `Dockerfile.base` used to carry a perl RUN block that stripped it. **That workaround is gone** — upstream fixed it in **3.5.0** (issue #1728 / PR #1717, merged 2026-06-14; `diary_write` now advertises `"required": ["agent_name"]` and enforces entry/content at dispatch), so the block was deleted when the pin moved to 3.5.0. **3.6.0** (2026-07-17) was reviewed for schema changes before bumping: it is purely additive/reliability (secure `serve` remote mode, optional Milvus, atomic KG `supersede()`, mining exclusions) and touches no MCP tool schema. **Ongoing rule:** before bumping this pin, diff the release notes for anything touching MCP tool schemas — that is the regression class this pin exists to catch — and bump **in lockstep with pi-devbox's `MEMPALACE_VERSION`**. NOTE: `MEMPALACE_VERSION` (the PyPI pin) and `MEMPALACE_TOOLKIT_REF` (the git ref for the `mempalace-toolkit` clone) are unrelated despite the shared prefix; do not conflate them.
 - **MemPalace install path** — installed via `uv tool install` into `/opt/uv-tools/mempalace/`. Both the `mempalace` CLI and the `mempalace-mcp` MCP server binary are shipped as entry points by the mempalace package itself and placed on PATH by uv as shims whose shebangs point at the venv's Python. No hand-rolled wrapper is needed. Do not use `pip install --break-system-packages` — that was the previous approach and has been removed. Do not use `["python3", "-m", "mempalace.mcp_server"]` in `opencode.jsonc` — system Python can't import from the uv venv.
 - **generate-config.py idempotency** — the script MUST never overwrite an existing `opencode.jsonc` or legacy `opencode.json`. Config persists in the `devbox-opencode-config` named volume; accidentally clobbering that file would destroy hand-edits. The smoke test asserts this.
 - **Skillset auto-deploy** — on every container start, `entrypoint-user.sh` looks for a skillset repo (detection order: `$SKILLSET_CONTAINER_PATH` → `$HOME/skillset` → `/workspace/skillset`) and runs `deploy-skills.sh --bootstrap --prune-stale`. This creates relative symlinks in `~/.agents/skills/` and `~/.config/opencode/instructions/`. Do NOT bind-mount `~/.agents/skills/` from the host — the container manages its own skills with relative symlinks that differ from the host's. The named volume `devbox-opencode-config` persists the deployed config across restarts.
-- **Config persistence via named volume** — `devbox-opencode-config` is a Docker named volume mounted at `~/.config/opencode/`. It is NOT a host bind mount by default. This separation allows both native and containerized opencode to coexist on the same machine without symlink conflicts. Users who need to override can replace the named volume with a host bind mount in their compose file. **Same pattern for pi:** `devbox-pi-config` is mounted at `~/.pi/` and persists user toggles (`/ext`-disabled extensions), `~/.pi/agent/settings.json` edits, and — because `NPM_CONFIG_PREFIX` is set to `~/.pi/npm-global` — anything installed via `pi install npm:...` or `npm install -g` as the developer user, across container recreate AND image rebuild.
-- **pi install contract** — `INSTALL_PI=true` (default false) opt-in build arg. The baked `pi` binary is npm-installed globally to `/usr` at build time (system prefix). At runtime, `NPM_CONFIG_PREFIX=/home/developer/.pi/npm-global` is set in the image ENV with that prefix's `bin/` prepended to `PATH` — so any `pi install npm:...` or `npm install -g` invoked by the developer user lands on the named volume and survives everything except `docker compose down -v`. The new ENVs are declared *after* all build-time `npm install -g` calls in the Dockerfile so they don't redirect the baked installs into a path that the volume mount would later shadow. If the user runs `npm install -g @earendil-works/pi-coding-agent` themselves, the user-installed copy on the volume wins via `PATH` order; otherwise image rebuild is the upgrade path for the baked pi (same contract as `OPENCODE_VERSION`). The pi-toolkit and pi-extensions repos are git-cloned into `/opt/` at build time, then their `install.sh` runs from `entrypoint-user.sh` on each container start to symlink into `~/.pi/agent/` (which lives on the named volume). The mempalace pi-bridge is symlinked manually from `/opt/mempalace-toolkit/extensions/pi/mempalace.ts` — we do NOT call mempalace-toolkit's full `install.sh` because its `install_skill` step would race with skillset auto-deploy `--prune-stale`.
-- **Pi deploy ordering matters in entrypoint-user.sh** — `pi-toolkit` runs first (creates `keybindings.json` symlink and writes pi-env.zsh), then `pi-extensions`, then `settings.json` template bootstrap, then mempalace bridge symlink. mempalace-toolkit's `check_pi_toolkit` probe (when called from the host install path) expects keybindings to already be present — not currently called from container, but ordering matches host convention.
-- **Default CMD is `bash -l`** — not a harness. `docker compose run --rm devbox` drops the user into a login shell to choose: `aws sso login`, then `opencode` or `pi` (or any tool). Pass the harness explicitly to launch directly: `docker compose run --rm devbox opencode` / `docker compose run --rm devbox pi`. `docker compose exec` bypasses entrypoint+CMD entirely (existing user workflow unchanged).
+- **OMOS bundled-skills reconcile** — on the omos variant, `entrypoint-user.sh` symlinks the five skills bundled with `oh-my-opencode-slim` (`clonedeps`, `codemap`, `deepwork`, `oh-my-opencode-slim`, `simplify`) from the image path `/usr/lib/node_modules/oh-my-opencode-slim/src/skills/<name>` into `~/.agents/skills/`, on **every** start, *after* the skillset deploy (so OMOS wins name collisions via `ln -sfn` — the only overlap is `simplify`, which was removed from the skillset repo). These are **absolute** symlinks (target is image-internal at a fixed `/usr` path) — do NOT "fix" them to relative like skillset's. Because the target lives in the image, pulling a newer image updates the skills with no installer run and no config reset. The block is non-fatal (`{ … } || true`), gated by `OMOS_SKILLS` (default true, **independent of `ENABLE_OMOS`**) and the presence of the source dir (no-op on the base variant). The two `oh-my-opencode-slim install` calls now pass `--skills=no` unconditionally — the installer manages only `oh-my-opencode-slim.json`, never skills; do not reintroduce installer-managed skills. A one-time migration (marker: `~/.config/opencode/.omos-skills-migrated`) backs up — never deletes — any frozen real copies the old installer left in `~/.config/opencode/skills/` to `<name>.bak.<epoch>`, because those would otherwise shadow the fresh image-sourced symlinks. The build-time smoke test asserts the bundled-skills source path exists (catches an upstream package restructure loudly). Full rationale: `docs/omos-skills.md`.
+- **Config persistence via named volume** — `devbox-opencode-config` is a Docker named volume mounted at `~/.config/opencode/`. It is NOT a host bind mount by default. This separation allows both native and containerized opencode to coexist on the same machine without symlink conflicts. Users who need to override can replace the named volume with a host bind mount in their compose file. Because `NPM_CONFIG_PREFIX` is set to `~/.config/opencode/npm-global` (relocated from the legacy `~/.pi/npm-global` in v2.0.0), anything installed via `npm install -g` as the developer user also lands on this volume and survives container recreate AND image rebuild.
+- **npm-global prefix relocation (v2.0.0 breaking change)** — the user-writable global npm prefix moved from `~/.pi/npm-global` to `~/.config/opencode/npm-global`. The old path lived on the `devbox-pi-config` volume (only mounted in `docker-compose.yml`); the new path is on `devbox-opencode-config`, which is a persistent named volume in BOTH `docker-compose.yml` and `docker-compose.shared.yml`. `entrypoint-user.sh` carries a one-time migration shim: if `~/.pi/npm-global` exists and the marker `~/.config/opencode/npm-global/.migrated-from-dot-pi` is absent, it `cp -an` the old `lib/`/`bin/`/`share/` into the new prefix (never overwriting fresh installs) and writes the marker. Baked binaries stay on `/usr` (the variant Dockerfile runs each `npm install -g` with `NPM_CONFIG_PREFIX=/usr`) so the volume mount doesn't shadow them. The `ENV NPM_CONFIG_PREFIX`/`PATH` lines in `Dockerfile.base` are declared *after* all build-time installs.
+- **Default CMD is `bash -l`** — not a harness. `docker compose run --rm devbox` drops the user into a login shell to choose: `aws sso login`, then `opencode` (or any tool). Pass the harness explicitly to launch directly: `docker compose run --rm devbox opencode`. `docker compose exec` bypasses entrypoint+CMD entirely (existing user workflow unchanged).
 - **Docker Hub description update** — uses `/v2/auth/token` endpoint (not the deprecated `/v2/users/login`). Auth uses `identifier`/`secret` fields, returns `access_token`, sent as `Bearer`. Short description must be ≤100 bytes.
 
 ## CI quirks
 
 - Both build jobs include an IPv4 preference step (`gai.conf` + `driver-opts: network=host` for buildx) to work around intermittent IPv6 failures on the Gitea runners.
-- `update-description` job runs only when both builds succeed (`needs: [build-base, build-omos]`).
-- Tags must be pushed to trigger the publish workflow. The validate workflow runs on push to main and PRs.
+- `update-description` job runs when the base variant published (`needs: [build-variant-base, build-variant-omos]`, gated with `always()` + an explicit `build-variant-base.result == 'success'` check so a partial-publish run still refreshes the Hub description).
+- **No image build happens on a push — builds are tag-triggered or explicitly requested.** The three workflows divide cleanly: `lint.yml` is cheap checks on every push/PR (actionlint + shell guard, hadolint, `docs-check`) and is the **only** workflow a push to main triggers; `validate.yml` is the amd64 build + smoke test and has **no push trigger** (pull_request / workflow_dispatch only); `docker-publish-split.yml` is the release path (tag-only). So pushing work-in-progress to main is free. It is safe because the release path already fails closed: `docker-publish-split.yml` pushes variant tags only after `smoke-base`/`smoke-omos` pass and promotes `base-latest` last, so an aborted release leaves at worst an unreferenced `base-<hash>` blob on Hub — never a half-published version tag. `validate-base`/`validate-omos` additionally keep a now-redundant `github.event_name != 'push'` clause as belt-and-braces, so re-adding a push trigger can't silently re-enable builds. **To validate before tagging**, open a PR, `workflow_dispatch` the Validate workflow, or `workflow_dispatch` `docker-publish-split.yml` against a throwaway tag with `promote_latest=false` (the only option that also exercises a changed base, which `validate.yml` structurally cannot — it builds variants on top of Hub's `base-latest`). **Keep cheap checks in `lint.yml`, not `validate.yml`** — anything that needs no image belongs in the workflow that actually runs on push; that is why `docs-check` lives there. If you change this model, update this bullet, both workflow header comments, and the two *File roles* entries together.
+- Tags must be pushed to trigger the publish workflow. A push to main triggers `lint.yml` only.
 - Smoke tests run on amd64 only (single-arch load into the local daemon). The multi-arch push happens after smoke passes.
-- **Gitea Actions runner has ~40 GB disk, often 70%+ used at job start.** All ten `load: true` jobs (`validate-base`, `validate-omos`, `validate-with-pi`, `validate-omos-with-pi`, `validate-pi-only`, `smoke-base`, `smoke-omos`, `smoke-with-pi`, `smoke-omos-with-pi`, `smoke-pi-only`) include a `Reclaim runner disk` step that strips catthehacker-resident toolchains and prunes stale docker state before `setup-buildx-action`. Build jobs use a lighter version (push-by-digest doesn't need `docker system prune`). Don't remove these steps without testing on a fresh runner.
+- **Gitea Actions runner has ~40 GB disk, often 70%+ used at job start.** All `load: true` jobs (`validate-base`, `validate-omos`, `smoke-base`, `smoke-omos`) include a `Reclaim runner disk` step that strips catthehacker-resident toolchains and prunes stale docker state before `setup-buildx-action`. Build jobs use a lighter version (push-by-digest doesn't need `docker system prune`). Don't remove these steps without testing on a fresh runner.
 - **`docker/build-push-action@v7` with `platforms: linux/amd64,linux/arm64` handles multi-arch push natively in a single job** — produces a proper manifest list, no matrix or merge step needed. An earlier revision split into per-arch matrix jobs with digest artifacts, but that pattern requires `actions/{upload,download}-artifact@v4+` which Gitea Actions doesn't support (see below).
 - **`actions/upload-artifact` and `actions/download-artifact` must stay at @v3 on Gitea.** v4+ uses a GitHub-Enterprise-specific Artifact API; runs fail with `GHESNotSupportedError`. If you need artifacts for a new reason (build logs, SBOMs, etc.), pin @v3 explicitly.
 - **Step scripts run under `/bin/sh` (dash), not bash.** Avoid bash-isms like `${VAR//a/b}` parameter-pattern substitution; use POSIX alternatives (`tr`, `sed`) or declare `shell: bash` on the step.
 - **`BUILDKIT_PROGRESS=plain`** is set at workflow level on `docker-publish-split.yml` so arm64-under-QEMU builds log each layer line-by-line. The default collapsed progress UI hides which step is stalled, which made diagnosing earlier hangs expensive.
 
+## Gitea API access (env token)
+
+`GITEA_ACCESS_TOKEN` + `GITEA_HOST` are passed into the container from the
+host `.env` via `docker-compose.yml` / `docker-compose.shared.yml`
+(`${GITEA_ACCESS_TOKEN:-}` / `${GITEA_HOST:-}`), primarily to enable the
+`gitea-mcp` server (see `generate-config.py`). They are **not** baked into
+the image. When configured, they are also available for **any** direct
+Gitea API interaction from inside the container — inspecting CI runs,
+checking published tags, listing commits — e.g.
+`curl -H "Authorization: token $GITEA_ACCESS_TOKEN" "$GITEA_HOST/api/v1/repos/joakimp/opencode-devbox/actions/runs?limit=5"`.
+Prefer this over a short-lived PAT file when the env token is present (the
+`ci-release-watcher` skill auto-detects it). Public-repo GET listings work
+unauthenticated (see the `resolve-versions` mempalace-toolkit note above), so
+the token matters mainly for private repos or rate-limit headroom; its
+lifecycle is host-managed, so there is nothing to revoke after use. Never
+echo the token value (including into logs).
+
+**Gotcha — the jobs endpoint takes the internal `id`, NOT the `run_number` the
+UI shows as `#239`.** The two diverge widely (`id=534` was `run_number=238`),
+and `GET .../actions/runs/<run_number>/jobs` does **not** error — it silently
+returns the jobs of a *different* run. This cost a wrong conclusion once: while
+verifying that a push no longer triggers builds, querying `.../runs/238/jobs`
+returned `docs-check`/`validate-base`/`validate-omos`, making a `lint.yml` run
+look like it had built images. Always read `id` from the run listing and use
+that. Two reliable patterns:
+
+```bash
+# Authoritative: which runs did MY commit trigger? Filter on head_sha — do not
+# trust ordering or run numbering.
+curl -sS -H "Authorization: token $GITEA_ACCESS_TOKEN" \
+  "$GITEA_HOST/api/v1/repos/joakimp/opencode-devbox/actions/runs?limit=20" \
+  | jq --arg sha "$(git rev-parse HEAD)" \
+      '.workflow_runs[] | select(.head_sha==$sha) | {id, run_number, path, event, status, conclusion}'
+
+# Then the per-job breakdown, keyed by the id from above (NOT run_number)
+curl -sS -H "Authorization: token $GITEA_ACCESS_TOKEN" \
+  "$GITEA_HOST/api/v1/repos/joakimp/opencode-devbox/actions/runs/<id>/jobs" \
+  | jq '.jobs[] | {name, status, conclusion}'
+```
+
+Counting runs per commit is also the cheapest way to assert the trigger model
+from the previous section: commits before the split show **2** runs
+(`lint.yml` + `validate.yml`), commits after show **1** (`lint.yml`). Note
+`lint.yml` is cheap in CPU but not always in wall-clock — its `actionlint` job
+`apt-get install`s shellcheck inside the container and has taken 6–15 min on a
+busy runner, so a still-`in_progress` lint run is not evidence of a problem.
+
 ## Testing changes
 
 The smoke test (`scripts/smoke-test.sh`) is the canonical check and runs automatically in CI. To run locally:

CHANGELOG.md

@@ -2,13 +2,938 @@
 
 All notable changes to the opencode-devbox container image.
 
-Tags follow `v{opencode_version}[letter]` — bare tag for the first build on a new opencode release, letter suffix (`b`, `c`, …) for container-level rebuilds on the same version. See [AGENTS.md](AGENTS.md#versioning-scheme) for details.
+Tags follow **independent semver** (since `v2.0.0`) — they version *this image*, not the bundled opencode release. MAJOR = breaking run/config changes, MINOR = backward-compatible features, PATCH = opencode/tool version bumps and small fixes. Pre-`v2.0.0` tags used the older `v{opencode_version}[letter]` scheme. See [AGENTS.md](AGENTS.md#versioning-scheme) for details.
 
 ---
 
-## Unreleased
+## v2.9.0 — 2026-08-04
 
-_(no changes since v1.15.13c)_
+Minor release. Headline: **real-browser verification lands in the base** (`agent-browser` + Playwright Chromium), the build manifest finally has a **reader** (`opencode-devbox-version`), and defaults move to **`claude-opus-5` / `gpt-5.6`** alongside opencode `1.17.20 → 1.18.13` and mempalace `3.5.0 → 3.6.0`. One behaviour change to read before upgrading: **`yq` is now mikefarah's Go v4, not Debian's Python `yq`** (see *Changed*). Several changes touch `Dockerfile.base`/`rootfs/`, so the **base image rebuilds** this release.
+
+### Added
+
+- **`agent-browser` + Playwright Chromium in the base — the agent can now drive a real browser.** Ported from pi-devbox. Lets an agent open pages, click/fill, `eval` JavaScript, snapshot the DOM and take screenshots, so front-end work can be **verified** (live DOM, layout, popup positioning, WebGL) instead of assumed. `AGENT_BROWSER_EXECUTABLE_PATH` is preset to `/usr/local/bin/agent-chrome`, a stable symlink the Dockerfile resolves with `find` rather than hardcoding — Playwright's browser directory is per-version *and* per-arch (`chrome-linux` on arm64, `chrome-linux64` on amd64), and the headless-shell binary is named `chrome-headless-shell` so `-name chrome` skips it. Playwright's redundant `chromium_headless_shell-*` build is deleted and the apt/npm caches cleaned, trimming the layer to **~625 MB** from ~960 MB. New floated build args `AGENT_BROWSER_VERSION` and `PLAYWRIGHT_VERSION`. This is now the single largest thing in the image and ships in **both** variants — a deliberate tradeoff, since verification is broadly useful.
+- **`opencode-devbox-version` — the build manifest is no longer invisible.** The image has baked ground truth to `/etc/opencode-devbox/build-manifest.json` for several releases (release tag, build date, source commit, live `opencode --version`, installed omos version, `mempalace-toolkit` HEAD) but **nothing read it** and nothing printed it, so answering "which image am I running?" meant knowing the path by heart. The new command wraps it in three modes (default human summary, `--json` for scripting, `--quiet` for a one-line `tag (rev)` form) and `entrypoint-user.sh` prints it as its **first** output, before the setup noise. It also performs a **live-vs-baked drift check**: because `NPM_CONFIG_PREFIX` points at the persistent `devbox-opencode-config` volume, a user's `npm install -g opencode` can shadow the baked `/usr` binary — so the command reports the live version and flags a mismatch rather than trusting the manifest blindly.
+- **`ENV COLORTERM=truecolor`.** Completes a true-colour story the image already half-shipped (it had `ncurses-term` + `kitty-terminfo` + the compiled `xterm-ghostty` alias + a system-wide Neovim `termguicolors` default, but never advertised 24-bit capability), so colour-aware tools like `bat` and `delta` stop falling back to 256 colours. Override with `COLORTERM=` (empty) from a terminal without true-colour support.
+- **Smoke coverage for the new surfaces and for `release_tag`.** Adds assertions for `agent-browser --version`, that `agent-chrome` resolves to an *executable* (catches a Playwright layout change rather than merely a dangling symlink), `AGENT_BROWSER_EXECUTABLE_PATH`, `COLORTERM`, the manifest's `release_tag` field, and all three modes of `opencode-devbox-version`. The pre-existing *Build provenance* block (manifest present, component fields, and the `! grep -q '"unknown"'` unresolved-component guard) was already in place and is unchanged.
+
+### Changed
+
+- **CI: a push to `main` no longer builds an image — the three workflows now divide cleanly by cost.** `lint.yml` is the cheap-checks workflow and the **only** one a push to main triggers: actionlint + the Gitea shell guard, hadolint, and `docs-check` (the `DOCKER_HUB.md` sync check, **moved here from `validate.yml`** so it survives — anything that needs no image belongs in the workflow that actually runs on push). `validate.yml` keeps the amd64 build + smoke test but **lost its push trigger entirely**: `pull_request` and a new `workflow_dispatch` only. `docker-publish-split.yml` is unchanged, still tag-only. `validate-base`/`validate-omos` also keep a now-redundant `github.event_name != 'push'` clause as belt-and-braces, so re-adding a push trigger cannot silently re-enable builds. `lint.yml` was renamed `Lint workflows` → `Lint` since it now covers Dockerfiles and docs too (no references to the old name existed). This brings the repo in line with pi-devbox, where `lint.yml` is likewise the only push-triggered workflow. It is safe because the release path already fails closed — `docker-publish-split.yml` pushes variant tags only after `smoke-base`/`smoke-omos` pass and promotes `base-latest` last, so an aborted release leaves at worst an unreferenced `base-<hash>` blob on Hub, never a half-published version tag. The pre-tag safety net remains available three ways: open a PR, dispatch the Validate workflow, or dispatch `docker-publish-split.yml` against a throwaway tag with `promote_latest=false` (the only route that also exercises a **changed base**, which `validate.yml` structurally cannot — it builds variants on top of Hub's `base-latest`). Note the practical coverage lost is narrower than it looks: the build jobs were already skipped whenever a commit touched `Dockerfile.base`, `rootfs/`, or `entrypoint*.sh`, so they only ever ran for variant-only changes — most usefully a bare `OPENCODE_VERSION` bump, for which a dispatch before tagging is now the equivalent.
+- **opencode bumped `1.17.20` → `1.18.13`** (`Dockerfile.variant` `OPENCODE_VERSION`; latest stable on npm, verified with `npm view opencode-ai version`). Verified as a safe minor-line jump by diffing upstream source at the tags rather than only reading release notes: `packages/core/src/config.ts`, `packages/core/src/config/provider.ts` and `packages/core/schema.json` are **byte-identical** to 1.17.20, so `generate-config.py` needs no change; no breaking Core changes (all "Desktop" notes are the Electron app, irrelevant here); nothing in the repo parses opencode CLI output beyond `--version`. Provider/MCP work in the range is net-positive (restored legacy MCP SDK client compatibility in 1.18.9, fixed MCP SSE reconnect loops in 1.18.11, better MCP OAuth in 1.18.8). The `OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS` removal-trigger was re-checked **at source level** — `packages/opencode/src/effect/runtime-flags.ts` is unchanged across all three tags and still gates the flag behind `OPENCODE_EXPERIMENTAL` — so it has **not** fired and the ENV stays baked ON.
+  - **`1.18.12` → `1.18.13` (published mid-audit, 2026-08-04) re-verified separately.** 19 commits, and the 249-file compare payload is *under* GitHub's 300-file cap, so the change list is complete rather than truncated. Content is almost entirely the Electron desktop app and localisation — new `az`/`fi`/`hi`/`id`/`it`/`nl`/`pa`/`sv` locales, RTL support, `desktop-menu` native translations — plus a `fix(github): include pull request identity in context` touching only `packages/opencode/src/cli/cmd/github.handler.ts` (the `github` subcommand, unused here) and a revert of an unreleased "fix slow queries". Confirmed by **hashing the five contract-surface files at both tags**: `config.ts`, `config/provider.ts`, `schema.json`, `runtime-flags.ts` and `agent/subagent-permissions.ts` are all identical, so neither `generate-config.py` nor the subagent-permissions note below is affected.
+  - **One upstream behaviour change worth knowing:** opencode **1.18.2** stopped subagents launching *nested* subagents by default (`packages/opencode/src/agent/subagent-permissions.ts` now denies the `task` permission to a spawned subagent unless its own ruleset grants it). `oh-my-opencode-slim` already adapted in **2.2.3** ("remove redundant subagent depth limiting", released one day later) by dropping its own client-side depth limiting in favour of opencode's native mechanism. Since CI resolves omos to the current **2.2.9**, no action is required here — but this is the one change in the whole range with real behavioural teeth, so the omos variant's multi-agent flow is worth exercising once on this image.
+- **`yq` is now mikefarah's Go `yq` v4, not Debian's Python `yq` — behaviour change.** The apt package on Debian/Ubuntu is the unrelated kislyuk/`yq`, a **jq-syntax wrapper** on a 3.x version line; it is a *different program* that happens to share the command name. It has been dropped from the apt list and the mikefarah binary is installed from GitHub instead (multi-arch, following the repo's `latest` convention, pin with `--build-arg YQ_VERSION=vX.Y.Z`). Two reasons: the `cloud-init` repo's `provision.sh`/`deploy.sh` require v4 syntax, and `THIRD_PARTY.md` **already credited "yq (mikefarah)"** while the image shipped the Python one — so this also closes a documented-vs-shipped mismatch. Brings parity with pi-devbox (its v1.2.3). The smoke test now pins the contract with `yq --version | grep -qE 'mikefarah.*version v4'`, so both a regression to the apt package and a surprise future yq v5 fail CI loudly. **Action required only if** you have scripts in this image calling `yq` with jq-style syntax — they will need porting to v4 expressions.
+- **Default models bumped to the Opus tier.** In `rootfs/usr/local/lib/opencode-devbox/generate-config.py`: `DEFAULT_MODELS["anthropic"]` → `anthropic/claude-opus-5` (was `claude-sonnet-5`; also feeds `FALLBACK_MODEL`), `DEFAULT_MODELS["amazon-bedrock"]` → `amazon-bedrock/global.anthropic.claude-opus-5`, and `DEFAULT_MODELS["openai"]` → `openai/gpt-5.6` (was `gpt-5.4`, which had gone stale — `gpt-5.6` shipped 2026-07-09, four days *before* the v2.8.0 cut). `claude-opus-5` was released 2026-07-24, i.e. after v2.8.0. Takes effect only for **new** containers with no `OPENCODE_MODEL` override and no existing config — `generate-config.py` still never overwrites an existing `opencode.jsonc`, which lives on a persistent volume. `.env.example` updated to match.
+- **mempalace pinned version bumped `3.5.0` → `3.6.0`** (`Dockerfile.base`), in lockstep with pi-devbox as the pin's comment requires. 3.6.0 (2026-07-17) is additive/reliability — secure `mempalace serve` remote mode, optional Milvus backend, atomic KG `supersede()`, conversation chronology, mining exclusions, plus recovery/locking fixes. Reviewed for MCP tool-schema changes before bumping (that being the exact regression class this pin exists to catch): there are **none**, and nothing touches `diary_write`. Two fixes are directly relevant to how this image uses mempalace: read-only mode now covers `checkpoint` + `delete_by_source` in `_MUTATING_TOOLS` (#1930), and agent attribution is preserved in `mempalace_checkpoint` (#2023/#2034).
+- **Smoke-test size thresholds lifted +650 MB** — base `2950 → 3600`, omos `3650 → 4300` — for the agent-browser/Chromium layer, sized to preserve roughly the same ~250 MB of headroom the previous values had so the guardrail keeps catching *runaway* growth rather than tripping on routine apt drift. A note was added warning **not** to copy pi-devbox's threshold number across: it sums `docker history` while this repo uses `docker image inspect .Size`.
+- **Refreshed the bg-subagents removal-trigger "last-checked" markers** (`Dockerfile.base`, `AGENTS.md`) from `opencode 1.17.20 / omos 2.2.0` to `1.18.13 / 2.2.9`, now recording that the check was done against upstream source and not just changelog prose.
+
+### Documentation
+
+- **New README section "Choosing a provider and model"** — makes explicit that the baked defaults are *only* defaults and that nobody is locked to Anthropic or Bedrock: a per-provider default table, `OPENCODE_MODEL` override examples (including a non-Anthropic provider), and the three facts that actually trip people up — defaults seed only a *new* config, an existing `opencode.jsonc` on the persistent volume is never rewritten (so changing `OPENCODE_MODEL` later has no effect until you edit or delete it), and switching model needs no rebuild.
+- **New README section "Browser automation (agent-browser)"** with usage examples, the symlink/versioned-path rationale, and an explicit size note for fork maintainers who'd rather drop the layer.
+- **README: `opencode-devbox-version` documented** under *Build provenance* with sample output and the drift-check explanation; `COLORTERM` covered under *Terminal compatibility*; new `COLORTERM` and `AGENT_BROWSER_EXECUTABLE_PATH` rows in the env table; the `OPENCODE_MODEL` row now points at the new section.
+- **README *Build Args* table drift fixed — four missing args added.** `AGENT_BROWSER_VERSION`, `PLAYWRIGHT_VERSION` and `YQ_VERSION` for the new tools, plus **`GITLEAKS_VERSION`**, which had been missing from the floated-args row despite existing as an ARG. Also added rows for the two *pinned* args that were absent entirely, `MEMPALACE_VERSION` and `DEBIAN_VERSION`, and refreshed the stale `--build-arg OPENCODE_VERSION=1.17.20` example. This is the third consecutive audit to find drift in this one table (v2.8.0 caught `MICRO`/`TEALDEER`/`TYPST`), which is why AGENTS.md carries a standing reminder about it.
+- **AGENTS.md: the stale MemPalace `anyOf` convention rewritten.** It still described a perl RUN block that had already been **deleted** when the pin moved to 3.5.0, and asserted "mempalace PyPI latest is 3.4.0 (== our pin), no release contains the fix yet, the workaround must stay" — all three false. Replaced with an accurate account: the workaround is gone, upstream fixed it in 3.5.0, the pin is 3.6.0, and the *standing* rule is preserved (review release notes for MCP tool-schema changes before bumping, and bump in lockstep with pi-devbox).
+- **AGENTS.md: two new conventions** — the agent-browser/Chromium size coupling (thresholds must move in the same commit; don't copy pi-devbox's number; why the Chrome path is `find`-ed) and the `yq` identity trap (don't revert it to the apt list, and never install both, because with two `yq` binaries on PATH the meaning of `yq` depends silently on PATH order). Also a *File roles* entry for `rootfs/usr/local/bin/opencode-devbox-version`.
+- **THIRD_PARTY.md**: added `agent-browser`, Playwright, and Chromium. The existing "yq (mikefarah)" credit is now actually true.
+- Folds in the three docs-only commits made after the v2.8.0 tag, which never got their own `Unreleased` block: a stale AGENTS.md push-step count (5 → 2 variant), a macOS NFD-filename gotcha noted for `dscp`/`scp` in the baked skill, and a re-sync of the vendored mempalace skill snapshot from skillset `63f3bf5`.
+
+## v2.8.0 — 2026-07-13
+
+Minor release. Headline: **default models move to `claude-sonnet-5`** and opencode bumps `1.17.15 → 1.17.20`. Rounds out with a large user-docs backfill — five v2.4.0–v2.7.0 base features (typst PDF export, terminal terminfo, Neovim 24-bit colour, the first-shell host SSH reachability check, the baked global gitignore) that had shipped without README prose — plus the bg-subagents removal-trigger "last-checked" markers refreshed to current versions. The marker refresh edits `Dockerfile.base`, so `base-<hash>` advances and the **base image rebuilds** this release (not just the variants).
+
+### Changed
+
+- **Default models bumped to `claude-sonnet-5`.** In
+  `rootfs/usr/local/lib/opencode-devbox/generate-config.py`,
+  `DEFAULT_MODELS["anthropic"]` → `anthropic/claude-sonnet-5` (was
+  `claude-sonnet-4-6`; also feeds `FALLBACK_MODEL`) and
+  `DEFAULT_MODELS["amazon-bedrock"]` →
+  `amazon-bedrock/global.anthropic.claude-sonnet-5` (was
+  `...claude-sonnet-4-5-20250929-v1:0`). Also updates the commented
+  `OPENCODE_MODEL` example in `.env.example`. Takes effect for new containers
+  when no explicit `OPENCODE_MODEL` override is set. The `openai` default
+  (`gpt-5.4`) is unchanged.
+- **opencode bumped `1.17.15` → `1.17.20`** (`Dockerfile.variant`
+  `OPENCODE_VERSION`; latest stable on npm, verified via `npm view
+  opencode-ai version`). 1.17.16–1.17.20 are TUI/cosmetic and internal fixes
+  (command-palette flash, sub-agent task-row styling, an OpenAI Responses
+  workaround removal) — no config-affecting or breaking changes. The
+  `OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS` removal-trigger was re-checked
+  against these release notes and has **not** fired; the flag stays baked ON.
+- **Refreshed the bg-subagents removal-trigger "last-checked" markers** (`Dockerfile.base`, `AGENTS.md`) from `opencode 1.17.15 / omos 2.1.0` to `opencode 1.17.20 / omos 2.2.0`. Comment-only and the trigger has still **not** fired — but the `Dockerfile.base` edit changes the base-content hash, so `base-<hash>` advances and **the base image rebuilds** on the next release (not just the variant).
+
+### Documentation
+
+- **README backfill for v2.4.0–v2.7.0 features that shipped without user-facing docs:** PDF export via `pandoc --pdf-engine=typst` (new *Document conversion* section), terminal `TERM`/terminfo support (`ncurses-term` + `kitty-terminfo` + `xterm-ghostty`), Neovim 24-bit colour default (`termguicolors` via system `sysinit.vim`), the first-shell host SSH reachability check, and the baked global gitignore. Also adds the three floated build args missing from the Build Args table (`MICRO_VERSION`, `TEALDEER_VERSION`, `TYPST_VERSION`). Docs-only — no image change.
+- **README + AGENTS.md polish:** refreshed stale `OPENCODE_VERSION=` build-arg examples (`1.5.0`, `1.17.8` → the current `1.17.20`), linked the new `THIRD_PARTY.md` from the *License* section, and added an AGENTS.md reminder to add any new floated `*_VERSION` ARG to the README *Build Args* table (the coupling gap that hid `MICRO`/`TEALDEER`/`TYPST` for three releases).
+
+## v2.7.0 — 2026-07-13
+
+Minor release. Headline: **further parity with the sibling `pi-devbox` image** —
+ports the base additions from pi-devbox v1.4.0 + v1.5.0 that opencode-devbox
+lacked. **PDF export now works out of the box** (`typst` as the pandoc PDF
+engine — pandoc shipped since v2.6.0 as a front-end only), modern terminal
+emulators resolve their `TERM` over SSH (ncurses-term + kitty-terminfo + a
+compiled `xterm-ghostty` alias), Neovim renders in readable 24-bit colour by
+default, and a first-shell **host SSH reachability check** warns when the Mac
+host isn't reachable. Also adds repository hygiene (LICENSE, THIRD_PARTY.md,
+hadolint CI, IDEAS backlog) and a `.claude` gitignore-seed pattern. All the
+tooling/rootfs changes are base-affecting, so `base-<hash>` advances and the
+base rebuilds.
+
+### Added
+
+- **`typst` — PDF engine for pandoc (Markdown→PDF).** `pandoc` shipped in the
+  base since v2.6.0 but as a front-end only — with no PDF back-end,
+  `pandoc -o out.pdf` failed with "xelatex not found". The base now installs
+  `typst`, a single ~30 MB static Rust binary (no LaTeX), used via
+  `pandoc --pdf-engine=typst`; a fuller TeX Live remains the higher-fidelity
+  fallback (install on demand). Also patches pandoc's bundled typst template
+  (`/usr/share/pandoc/data/templates/template.typst`), whose empty `font: ()`
+  default otherwise made a naked `--pdf-engine=typst` fail with "font fallback
+  list must not be empty" — defaulted to `Libertinus Serif` so PDF export works
+  without `-V mainfont`. Adds `xz-utils` to the apt layer (typst ships a
+  `.tar.xz`). Tracks `latest`; pin with `--build-arg TYPST_VERSION=vX.Y.Z`.
+  Ported from pi-devbox v1.4.0 (+ its v1.5.0 font-default fix). Base-affecting.
+
+- **Terminal support beyond the default: `ncurses-term` + `kitty-terminfo` + a
+  compiled `xterm-ghostty` alias.** The base previously shipped only
+  `ncurses-base`, so SSHing in from a modern emulator degraded to a dumb
+  fallback. Now installs `ncurses-term` (terminfo for WezTerm, Alacritty, foot,
+  st, the base `ghostty` entry, and many more) and `kitty-terminfo`
+  (`xterm-kitty`), and compiles an `xterm-ghostty` alias with `tic -x`
+  (`use=ghostty`) — Ghostty connects as `TERM=xterm-ghostty` and no distro
+  packages that name. iTerm2's `xterm-256color` is already in ncurses-base. New
+  `rootfs/usr/local/share/terminfo-src/ghostty.terminfo`; the build asserts the
+  entry landed via `infocmp`. Ported from pi-devbox v1.5.0. Base-affecting.
+
+- **Readable Neovim colours out of the box.** New system-wide Neovim config
+  (`rootfs/etc/xdg/nvim/sysinit.vim`) enables `termguicolors`. Vanilla Neovim
+  otherwise fell back to a muddy, low-contrast 256-colour palette over ssh.
+  `sysinit.vim` is Neovim's system vimrc — it loads for every user before any
+  personal `~/.config/nvim` and can still be overridden per-user
+  (`:set notermguicolors`). Pairs with `kitty-terminfo` above. Ported from
+  pi-devbox v1.5.0. Base-affecting.
+
+- **Host SSH reachability check at shell startup.** `~/.bash_aliases` (baked
+  into the base) now runs a one-time SSH probe on the first bash session of each
+  container. If the Mac host is not reachable (Remote Login disabled or the
+  `devbox_jump` key not yet authorized) it prints a clear warning with the exact
+  two steps to fix it, including the container's public key inline. Subsequent
+  shells skip the check (a `/tmp` flag cleared on recreate); silent when SSH
+  works. Complements the existing `setup-lan-access.sh` key-generation message.
+  Ported from pi-devbox v1.4.0. Base-affecting.
+
+- **Seeded global gitignore now ignores `**/.claude/settings.local.json`.**
+  Claude Code's per-machine local settings file holds machine-specific
+  permissions and can carry credentials, so it should never be committed. The
+  seed (`rootfs/home/developer/.gitignore_global`, baked to `/etc/skel-devbox/`)
+  gains the pattern. Existing containers are unaffected (the seed is copied only
+  when `~/.gitignore_global` is absent). Base-affecting (rootfs COPY).
+
+- **Repository hygiene: `LICENSE`, `THIRD_PARTY.md`, `hadolint` CI, `IDEAS.md`.**
+  The repo declared MIT only in prose; it now ships an actual `LICENSE` file
+  (MIT, © Joakim Persson) plus `THIRD_PARTY.md` recording that the published
+  images bundle third-party software under its own terms (opencode, OMOS,
+  mempalace — MIT; gosu/uv/typst Apache-2.0; Debian packages under their own
+  licenses). A new `hadolint` job (pinned v2.14.0) in the lint workflow lints
+  `Dockerfile.base` and `Dockerfile.variant`; `.hadolint.yaml` grandfathers the
+  deliberate choices (unpinned apt/npm, `cd`-in-`RUN`, `SC2086` — mirroring the
+  existing shellcheck excludes) and fails on anything new at `warning`+.
+  `IDEAS.md` parks vetted-but-unscheduled follow-ups. Repo/CI only — not baked
+  into the image.
+
+## v2.6.0 — 2026-07-08
+
+Minor release. Headline: **CLI-toolset parity with the sibling `pi-devbox`
+image** (adds `yq`, `pandoc`, `graphviz`, `tealdeer`/`tldr`, and the `dot-watch`
+wrapper) plus **background subagents enabled by default**
+(`OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=true`), and an opencode bump
+`1.17.13 → 1.17.15`. Touches `Dockerfile.base` (apt list, new binary, new ENV)
+so the base image rebuilds (`base-<hash>` advances); the opencode bump rebuilds
+the variant layers.
+
+### Added
+
+- **CLI tooling parity with pi-devbox.** Added `yq` (YAML query/transform,
+  pairs with `jq`), `pandoc` (Markdown↔HTML/etc. document converter), and
+  `graphviz` (`dot` diagram rendering) via apt; `tealdeer` (the `tldr` command —
+  quick command examples) as a ~5 MB static musl binary via the floated
+  `TEALDEER_VERSION` arg; and the `dot-watch` wrapper (`rootfs/usr/local/bin/dot-watch`,
+  auto-renders a `.dot` file to PNG on save, graphviz-only). Deliberately did
+  **not** port `socat`, `imagemagick`, or the pi-studio-specific `studio-expose`
+  bridge. smoke-test asserts presence of all five.
+- **Background subagents enabled by default (`OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=true`).**
+  Baked as a base-level `ENV` in `Dockerfile.base`. opencode gates native
+  background subagents behind this experimental flag; OMOS V2+ makes background
+  orchestration its default workflow, so the omos variant was effectively
+  degraded without it. Overridable at runtime (`-e OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=false`).
+  Documented in README env table and `.env.example`; smoke-test asserts the var
+  is present in the image environment. Marked experimental upstream — carries a
+  removal-trigger note in `Dockerfile.base` + AGENTS.md for when opencode
+  promotes it out of experimental (no upstream roadmap date as of opencode
+  1.17.15 / omos 2.1.0).
+
+### Changed
+
+- **opencode `1.17.13` → `1.17.15`.** v1.17.14 adds a code-mode MCP adapter for
+  confined orchestration scripts, hides the `execute` tool unless code mode is
+  enabled, and **fixes paginated MCP tool catalogs losing tool metadata +
+  output-schema validation** (relevant — this image wires the mempalace,
+  gitea-mcp, and context7 MCP servers). v1.17.15 is bugfixes + desktop/TUI UX
+  (Z.ai context-overflow error classification, macOS Sequoia titlebar, model
+  picker). No breaking changes. Release notes:
+  <https://github.com/anomalyco/opencode/releases>.
+
+---
+
+## v2.5.0 — 2026-07-02
+
+Minor release. Headline: **optional shared/external MemPalace** —
+`generate-config.py` can now register mempalace as a `remote` MCP endpoint
+(`MEMPALACE_REMOTE_URL`, optional `MEMPALACE_REMOTE_TOKEN` → Bearer) so one
+palace can serve pi + opencode + native, instead of a per-container local
+server; adds `docker-compose.mempalace.yml` for the shared server. Also fixes a
+`validate.yml` false-red on base-changing commits (skips variant validation and
+corrects the `rootfs/` base-input detector regex). opencode stays `1.17.13`; the
+base image rebuilds (`rootfs/…/generate-config.py` changed), so the new config
+generator lands in `base-<hash>`.
+
+### Added
+
+- **Optional shared/external MemPalace via `MEMPALACE_REMOTE_URL`.** MemPalace
+  can now be pointed at one shared HTTP endpoint instead of each container
+  running its own local server. Set `MEMPALACE_REMOTE_URL=http://<host>:8765/mcp`
+  (optionally `MEMPALACE_REMOTE_TOKEN`, sent as `Authorization: Bearer`) in
+  `.env`; `generate-config.py` then registers the `mempalace` MCP server as a
+  `remote` endpoint instead of the local `mempalace-mcp` stdio command (no local
+  server is spawned, and the `devbox-palace` volume becomes irrelevant). Leaving
+  it unset keeps the default local-per-container palace. Mirrors the same env
+  contract as pi-devbox's `mempalace.ts` bridge, so one palace can serve pi +
+  opencode + native.
+- **`docker-compose.mempalace.yml`** — an optional shared MemPalace server
+  (`mempalace-mcp --transport http`) that several containers/harnesses can point
+  `MEMPALACE_REMOTE_URL` at. Binds to `127.0.0.1:8765` by default (the HTTP
+  transport is unauthenticated — keep it on a trusted network or behind a proxy).
+
+### Fixed
+
+- **`validate.yml` no longer reports a false failure on base-changing commits.**
+  `validate-base`/`validate-omos` build variants from the published
+  `base-latest`; when a commit changed base inputs *and* tightened
+  `smoke-test.sh` in lockstep (as v2.4.0's nano/micro did), those jobs
+  hard-failed against the still-stale `base-latest` until the release rebuilt it.
+  The `base-change-warning` job now exports a `base_changed` output and those two
+  jobs **skip** when base inputs changed, deferring full base validation to the
+  release path (`docker-publish-split.yml`). The scary red becomes a neutral skip
+  plus the existing warning. Also fixes the base-input detector itself: its
+  `^(…|rootfs/|…)$` regex anchored the whole alternation, so the `rootfs/` branch
+  only matched a file literally named `rootfs/` and never real paths like
+  `rootfs/usr/…/generate-config.py` — `rootfs/` is now a prefix match, so
+  rootfs-only base changes are detected too.
+
+---
+
+## v2.4.0 — 2026-07-01
+
+Minor release. Adds two **non-modal editors** (`nano` + `micro`) alongside
+nvim, ports **CI hardening** from pi-devbox (the sh-vs-bash workflow guard and
+the `base-latest` digest-based promote fix), bakes a **global gitignore** into
+the image, and bumps **opencode `1.17.10` → `1.17.13`**. Because
+`Dockerfile.base` changed (nano + micro), this release rebuilds the base image.
+
+### Changed
+
+- **opencode `1.17.10` → `1.17.13`** (three upstream patches). Highlights:
+  session snapshots + revert controls (1.17.11); MCP OAuth reconnect/refresh
+  fixes, a TUI yolo auto-approve mode, and better default small models
+  (1.17.12); forced reasoning mode for OpenAI-compatible reasoning models plus
+  a GitHub Copilot stale-response-ID fix (1.17.13). Full notes:
+  <https://github.com/anomalyco/opencode/releases>.
+
+### Added
+
+- **Global gitignore baked into the image.** A `~/.gitignore_global`
+  (`*.bak`, `*.bak.*`, `*~`, `*.orig`, `*.swp`, `*.tmp`) is seeded into the home
+  dir from `/etc/skel-devbox/` on first boot (seed-if-absent, like
+  `.bash_aliases`/`.inputrc`, so user edits survive recreate) and wired via
+  `git config --global core.excludesFile`. Personal/tooling backup artifacts are
+  now ignored across all repos in the container without per-repo `.gitignore`
+  entries. The `core.excludesFile` wiring is skipped if the user already set one.
+
+- **Non-modal editors `nano` + `micro` alongside `nvim`.** The image shipped
+  only nvim (`EDITOR=nvim`), a modal vi-style editor. Added both a classic and
+  a modern non-modal option for users who don't want vi keybindings:
+  - **nano** (apt): ~2.8 MB installed; its deps (libc6, libncursesw6,
+    libtinfo6) are already present via nvim/less/htop/tmux, so no extra
+    packages are pulled in.
+  - **micro**: ~12 MB single static Go binary from GitHub releases (same
+    pattern as bat/eza/zoxide). Desktop-style keys (Ctrl+S/Ctrl+Q), mouse,
+    syntax highlighting. `ARG MICRO_VERSION` pins; defaults to latest.
+
+  Combined ~15 MB (<0.5% of the image). `EDITOR` stays `nvim`; both are opt-in
+  (`export EDITOR=micro | nano`). Uses the canonical `micro-editor/micro` URL
+  because the old `zyedidia/micro` org rename makes `/releases/latest` redirect
+  to another `/latest`, defeating the tag-parsing latest-resolution idiom.
+  Base-image change, so it lands on the next `base-<hash>` rebuild. Ported from
+  pi-devbox 3a59e15.
+
+- **Workflow-lint guard (`.gitea/workflows/lint.yml` + `scripts/check-workflow-shell.sh`).**
+  New cheap (~10s) lint workflow that runs on every push/PR (not just release
+  tags): a Gitea-accurate shell guard plus pinned `actionlint` + `shellcheck`.
+  The custom guard asserts every `run:` step in every `.gitea/workflows/*.yml`
+  resolves to an effective shell of `bash`, closing the actionlint blind spot
+  (actionlint models GitHub, whose default `run` shell is bash, so it does not
+  flag bash syntax in a step that merely omits `shell:` — the exact way the
+  sh-vs-bash footgun manifests on Gitea, whose default is `sh`/dash). Ported
+  from pi-devbox.
+
+### Changed (CI)
+
+- **Workflow-wide `defaults: run: shell: bash`** added to
+  `docker-publish-split.yml` and `validate.yml`. Gitea Actions' default step
+  shell is `sh` (dash), so bash-only syntax (`set -o pipefail`, `[[ ]]`,
+  arrays) in a step that forgets `shell: bash` fails silently. Setting the
+  default workflow-wide eliminates the whole class. All pre-existing steps use
+  only POSIX syntax, so bash (a superset) runs them unchanged — no behavioural
+  change. Preventive port from pi-devbox, where this class bit twice.
+
+### Fixed (CI)
+
+- **`promote-base-latest` re-points `base-latest` by digest, not `need_build`.**
+  The job gate keyed off `need_build == 'true'`, assuming `need_build == false`
+  meant `base-latest` was already current. A dry-run dispatch
+  (`promote_latest=false`) that pre-builds `base-<hash>` falsifies that: the
+  later tag run sees `need_build == false`, skips promotion, and leaves
+  `base-latest` one base behind. The gate now runs on every tag release /
+  promote dispatch, and the no-op optimization moved into the step as a `crane
+  digest` compare — it re-tags only when `base-latest` actually differs from the
+  released `base-<hash>` (genuine cache-hit releases stay a no-op). Workflow-only
+  change; base hash unaffected (no base rebuild). Ported from pi-devbox b7197e8
+  (which hit this on its v1.2.3 release, 2026-06-27).
+
+---
+
+## v2.3.0 — 2026-06-25
+
+Minor release. Adds an **image-baked fallback skills + harness-instruction**
+mechanism (ported from pi-devbox v1.2.0/v1.2.1, adapted to opencode's
+`instructions/` model), bumps **opencode `1.17.8` → `1.17.10`** and **mempalace
+`3.4.0` → `3.5.0`** (dropping the now-obsolete `diary_write` schema
+workaround), and ports the one outstanding doc fix from pi-devbox's SSH sidecar
+work. Defaults are unchanged when a skillset is mounted, so the canonical CI
+build differs only by the version bumps and the additive image layer.
+
+### Added: image-baked fallback skills + harness instruction
+
+The image now ships two skills and one harness instruction under
+`/usr/local/share/opencode-devbox/`, linked into place by `entrypoint-user.sh`
+on every start, so a container behaves correctly **even with no `skillset` repo
+mounted**:
+
+- **`opencode-devbox-environment`** (authored here) — the container-shaped facts
+  an agent needs: the three persistence tiers (and why image-owned content must
+  live under `/usr` rather than a home dir the `devbox-opencode-config` volume
+  shadows), the interactive-vs-tool-shell alias gotcha (`dssh`/`dscp`/`cat`→`bat`
+  only exist in interactive bash), host + LAN SSH reachability via the
+  `~/.ssh-local` sidecar and ControlMaster, split-horizon DNS, uv-first Python,
+  and the OMOS variant. Adapted from pi-devbox's `pi-devbox-environment`, minus
+  the pi-only pieces (pi-studio, fork/recall).
+- **`mempalace`** — a vendored snapshot of the skillset's consumer skill
+  (memory-continuity protocol). No `pi-extensions` skill is carried over —
+  opencode has no `fork`/`recall` extensions.
+- **`instructions/opencode-devbox.md`** — symlinked into
+  `~/.config/opencode/instructions/`, which opencode auto-loads as a
+  session-start system prompt. It proactively points the agent at the
+  `opencode-devbox-environment` skill and the MemPalace continuity protocol, so
+  a fresh container picks them up rather than relying on description-matching.
+
+Precedence is preserved: skills link **only-when-absent** (a mounted skillset or
+an OMOS-bundled skill of the same name always wins), and the instruction uses a
+distinct filename so it never collides with a mounted skillset's instructions.
+Link targets live in the image, so `docker compose pull` + recreate refreshes
+them for free — nothing is copied into the persistent config volume (which would
+freeze it; cf. the OMOS-skills history in `docs/omos-skills.md`). Build-time
+(`smoke-test.sh`) and runtime (`recreate-sanity-check.sh`) assertions cover both
+the baked source and the resolved links. See
+`rootfs/usr/local/share/opencode-devbox/skills/VENDORED.md`.
+
+### Changed
+
+- **opencode `1.17.8` → `1.17.10`.** Highlights across 1.17.9–1.17.10: MCP
+  server instructions are now added to session context; MCP resource template
+  listing + resource read tools; a new `--mini` CLI mode; skill base
+  directories emitted as filesystem paths instead of `file://` URLs; assorted
+  MCP/OAuth and agent-step-limit bugfixes. (Source: `anomalyco/opencode`
+  releases.)
+- **mempalace `3.4.0` → `3.5.0`** (lockstep with pi-devbox v1.2.2). 3.5.0 ships
+  the upstream fix for the top-level-`anyOf` `diary_write` schema (issue #1728 /
+  PR #1717, merged 2026-06-14): the advertised schema is now
+  `"required": ["agent_name"]` with entry/content enforced at dispatch, which
+  the Anthropic tools API accepts.
+
+### Removed
+
+- **The `diary_write` top-level-`anyOf` workaround in `Dockerfile.base`.** The
+  `perl` patch of the installed `mcp_server.py` is gone now that 3.5.0 fixes it
+  at the source (verified against the published wheel).
+
+### Fixed
+
+- **Stale `ssh-lan.conf` guidance comment** in `setup-lan-access.sh`. The
+  `INCLUDE_BLOCK` comment previously implied LAN-peer `ProxyJump` overrides go
+  in `~/.ssh/config` (typically bind-mounted read-only); corrected to point at
+  the host-owned `~/.config/devbox-shell/ssh-lan.conf` (mirrors pi-devbox
+  `8de0fad`). Comment-only; no behavior change.
+
+---
+
+## v2.2.0 — 2026-06-19
+
+Ports the build-provenance, CI-hardening, SSH and shell fixes that landed in
+the sibling **pi-devbox** repo (v1.1.4–v1.1.6) into opencode-devbox, adapted to
+this image's companions and two-variant (`base`/`omos`) shape. Also bumps
+opencode. Defaults are unchanged, so the canonical CI build stays byte-identical
+apart from the opencode bump and the (cache-free) provenance layer.
+
+### Fixed: read-only `~/.ssh` ControlPath / LAN sidecar on native Linux
+
+`rootfs/usr/local/lib/opencode-devbox/setup-lan-access.sh` previously
+`exit 0`-ed early on native-Linux hosts (`auto` mode, not VM-backed) **before**
+rendering the writable `~/.ssh-local/config` sidecar. On such hosts with a
+read-only `~/.ssh` bind-mount, `dssh`/`dscp` got no config and the `Host *`
+ControlPath redirect into `~/.ssh-local/cm` never happened, so a user
+`~/.ssh/config` carrying the CGNAT idiom `ControlPath ~/.ssh/cm/%r@%h:%p`
+broke ControlMaster. The sidecar (ControlPath redirect + `Include
+~/.ssh/config`) is now rendered on **every** host OS; only the jump-specific
+blocks (host alias, key generation, peer overrides, RFC1918 catch-all) stay
+gated behind a new `NEED_JUMP` flag. `Dockerfile.base` and `entrypoint-user.sh`
+comments updated to document the always-render behavior and the
+plain-`ssh <host>` caveat. (Mirrors pi-devbox v1.1.5; the pi-only
+`ssh-controlmaster` extension layer has no opencode equivalent and is N/A.)
+
+### Fixed: bash history loss in nested / tmux shells
+
+`rootfs/home/developer/.bash_aliases` exported the `DEVBOX_HIST_SET` flush
+guard, so it leaked into child processes — every nested shell (crucially each
+tmux pane, which inherits the tmux server's env) saw the guard already set and
+skipped installing `history -a` in `PROMPT_COMMAND`. Those shells only
+persisted history on a clean exit, silently losing in-memory history on abrupt
+termination (`docker stop`, `tmux kill-server`, SIGKILL). The guard is now
+shell-local (dropped `export`). (Mirrors pi-devbox v1.1.4.)
+
+### Added: build provenance — OCI labels + on-disk manifest
+
+The variant build now bakes OCI labels
+(`org.opencontainers.image.{version,revision,created}` +
+`se.jordbo.opencode-devbox.{opencode-version,install-omos,omos-version,mempalace-toolkit-ref}`)
+and writes `/etc/opencode-devbox/build-manifest.json` from **ground truth** —
+the live `opencode --version`, the installed `oh-my-opencode-slim` version
+(JSON `null` in the `base` variant), and the actual checked-out HEAD of
+`/opt/mempalace-toolkit` — so a published tag is self-describing and
+reconstructable after CI logs rotate. Provenance ARGs (`RELEASE_TAG`,
+`BUILD_DATE`, `SOURCE_REVISION`, re-declared `MEMPALACE_TOOLKIT_REF`) are
+declared last in `Dockerfile.variant` so they never bust the expensive
+npm-install layers. Wired into both `build-variant-*` and `smoke-*` jobs;
+`scripts/smoke-test.sh` now asserts the manifest exists, is complete, has no
+`unknown` components, and that the `opencode-version` OCI label is present.
+(Mirrors pi-devbox v1.1.6.)
+
+### Added: base-rebuild hash guard (`scripts/check-base-hash.sh`)
+
+New CI guard (run first in the `base-decide` job) that fails the build if any
+floating `ARG *_REF` consumed by `Dockerfile.base` is not folded into the
+`base_tag` hash — preventing the v1.1.2-class staleness footgun where a
+ref-only dependency change silently fails to rebuild the base. Passes today
+(`MEMPALACE_TOOLKIT_REF` is already folded in); this is forward protection.
+(Mirrors pi-devbox v1.1.6.)
+
+### Changed: fail-loud version/ref resolution
+
+The `resolve-versions` step now validates each resolved value — the
+mempalace-toolkit ref must be a 40-hex commit SHA, the omos version must be
+semver — and **aborts the release** on failure instead of silently falling
+back to a floating `main` ref (which defeats both cache-busting and
+reproducibility). The step also gains `shell: bash`, because `set -o pipefail`
+is illegal under the runner's default dash/sh and would otherwise abort the
+step (this exact latent bug bit pi-devbox's first v1.1.6 run). (Mirrors
+pi-devbox v1.1.6.)
+
+### Added: overridable `MEMPALACE_TOOLKIT_REPO` build-arg
+
+`Dockerfile.base` no longer hardcodes the mempalace-toolkit clone URL inline;
+it is now an `ARG MEMPALACE_TOOLKIT_REPO` defaulting to the canonical gitea
+origin, so a relocated/forked build can repoint it via `--build-arg` without
+editing the Dockerfile. Default unchanged. (Mirrors pi-devbox v1.1.6.)
+
+### Bumped: opencode-ai 1.17.7 → 1.17.8
+
+`OPENCODE_VERSION` ARG in `Dockerfile.variant`. `1.17.8` is the current npm
+`latest` stable. Only the variant layer rebuilds; the base is unaffected.
+
+### Added: opencode.json merge-on-recreate — non-destructive `.proposed` sidecar
+
+The pi-devbox v1.1.4 deep-merge into a preserved `settings.json` does not port
+cleanly here: opencode's config is *generated from env vars* and written as
+JSONC with comments (not a static image-owned template), and overwriting or
+`jq`-merging a possibly-bind-mounted host config is destructive. Instead,
+`generate-config.py` keeps its "never touch an existing config" guarantee and
+adds a non-destructive side-channel: when a live config exists, it writes
+`opencode.jsonc.proposed` — the config it *would* generate for the current
+environment plus this image's defaults — **only when that differs** from the
+live config, and removes it once they match. opencode never loads a `.proposed`
+file, so it is purely a manual-merge reference (e.g. surfacing a default MCP
+server added in a newer image). A one-line hint is logged when one is written;
+an unparseable live config surfaces the proposal rather than guessing. The
+proposed config is regenerated from env + image defaults, so a diff may reflect
+your own past edits as well as new image defaults — the file header says so.
+Covered by a new `scripts/smoke-test.sh` assertion (write-on-diff, removal on
+match, live config never clobbered).
+
+---
+
+## v2.1.2 — 2026-06-16
+
+Image-semver **patch**: bumps opencode to `1.17.7`. No devbox-side changes
+beyond the `OPENCODE_VERSION` ARG, so only the variant layer is rebuilt; the
+base is unaffected by this change.
+
+### Bumped: opencode-ai 1.17.6 → 1.17.7
+
+`OPENCODE_VERSION` ARG in `Dockerfile.variant`. Upstream `1.17.7` (published
+2026-06-14) is a bugfix-and-minor-improvement patch with no breaking,
+runtime-dependency, bundled-Bun, or CPU/AVX changes — a pure version bump on
+the devbox side. Upstream highlights:
+
+- **Core (bugfixes):** plugin client requests now reuse the active server
+  instead of assuming the default local port; ACP shell tool calls show the
+  command and working directory from the start; plugin-provided shell
+  environment variables now apply to PTY sessions.
+- **Core (improvements):** MCP servers can now receive the current workspace as
+  a client root.
+- **TUI:** MCP debug now uses the SDK's latest protocol version.
+- **Desktop:** the new-session route stays scoped to its own draft server, so
+  prompts and state target the right workspace.
+- **SDK:** clients refresh model and provider availability when integrations
+  change; credential update and remove calls accept `location`.
+
+Full notes:
+<https://github.com/anomalyco/opencode/releases/tag/v1.17.7>.
+
+## v2.1.1 — 2026-06-14
+
+Image-semver **patch**: bumps opencode and lands the `mempalace-toolkit`
+SHA-resolution CI fix plus two doc corrections that accumulated on `main` since
+`v2.1.0`. The toolkit change folds a live SHA into the base-tag hash, so this
+release carries a full base rebuild and a `base-latest` re-promote.
+
+### Bumped: opencode-ai 1.17.5 → 1.17.6
+
+`OPENCODE_VERSION` ARG in `Dockerfile.variant`. Upstream `1.17.6` (published
+2026-06-13) is a single Core bugfix — *"Improved MCP server compatibility by
+declaring OpenCode's supported client capabilities"* — with no breaking,
+runtime-dependency, bundled-Bun, or CPU/AVX changes. Pure version bump on the
+devbox side. Full notes:
+<https://github.com/anomalyco/opencode/releases/tag/v1.17.6>.
+
+### Changed
+
+- **`mempalace-toolkit` is now CI-resolved to a commit SHA.** It is the only
+  dependency cloned in `Dockerfile.base` (everything else is in the variant),
+  so it bypassed the `resolve-versions` → build-arg plumbing and its ref stayed
+  a literal `main` — meaning a toolkit-only fix would silently fail to land
+  unless `Dockerfile.base` itself changed. Now `resolve-versions` resolves the
+  `mempalace-toolkit` `main` HEAD to a SHA (new `mempalace_toolkit_ref` output,
+  via the gitea commits API — unauthenticated, no secret needed), `base-decide`
+  folds that SHA into the base-tag hash (so a moved toolkit forces a base
+  rebuild) and now **depends on `resolve-versions`**, and `build-base` passes it
+  as `--build-arg MEMPALACE_TOOLKIT_REF`. The base clone switched from
+  `git clone --branch` to a SHA-capable `git fetch <ref> + checkout FETCH_HEAD`.
+  Trade-off: `base_tag` now reflects a live gitea lookup — an API blip falls
+  back to `main` and triggers one *extra* rebuild, never a *missed* one. Updated
+  `.gitea/README.md` Step 1 and `AGENTS.md` Critical conventions in lockstep.
+
+### Docs (no image change)
+
+- Correct the MemPalace `diary_write` anyOf workaround watch-target: upstream
+  PR #1735 was **closed unmerged** (2026-06-11), so the old “remove once #1735
+  ships” TODO pointed at a dead PR. Issue #1728 is still open; PR #1717 is the
+  current live candidate; mempalace PyPI latest is still 3.4.0 (== our pin), so
+  the workaround stays. Rewrote the `Dockerfile.base` tracking comment and added
+  a durable note under `AGENTS.md` Critical conventions.
+
+- Fix the quick-start description in `README.md` and the Hub `HUB_TEMPLATE`
+  (`scripts/generate-dockerhub-md.py`, regenerated `DOCKER_HUB.md`): bare
+  `docker compose run --rm devbox` lands in a **login shell** (default `CMD` is
+  `bash -l`), not opencode. The old copy claimed the opposite and had a garbled
+  "Use `bash` instead of (no command)" half-sentence. Pass `opencode` explicitly
+  to start the harness directly. Doc-only — does not trigger a new image build.
+
+## v2.1.0 — 2026-06-13
+
+Image-semver **minor**: adds the OMOS bundled-skills image-symlink mechanism and
+bumps opencode. `v2.0.0` decoupled image versioning from the opencode version
+(see [AGENTS.md](AGENTS.md#versioning-scheme)); this is the first feature release
+on the semver line, mirroring how the sibling `pi-devbox` repo moved to semver
+after its own `v1.0.0` decouple.
+
+### Bumped: opencode-ai 1.17.4 → 1.17.5
+
+`OPENCODE_VERSION` ARG in `Dockerfile.variant`.
+
+### Changed: OMOS bundled skills are now symlinked from the image, not copied
+
+The five skills bundled with `oh-my-opencode-slim` (`clonedeps`, `codemap`,
+`deepwork`, `oh-my-opencode-slim`, `simplify`) are now **symlinked from the
+image** into `~/.agents/skills/` on every container start, instead of being
+**copied** into `~/.config/opencode/skills/` once on first run by the OMOS
+installer.
+
+Why: the old copy landed in the persistent `devbox-opencode-config` volume and
+was gated by config-existence, so it **froze** at whatever the image shipped on
+first run — pulling a newer image never refreshed the skills, and the only
+update path (`OMOS_RESET=true`) also clobbered the user's hand-tuned
+`opencode.jsonc`. With the symlink approach the link target lives in the image,
+so **`docker compose pull` + recreate updates the skills for free** — no
+installer run, no config reset.
+
+- `entrypoint-user.sh`: new OMOS bundled-skills reconcile block (runs after the
+  skillset deploy so OMOS wins the `simplify` name collision; absolute symlinks;
+  non-fatal; gated by `OMOS_SKILLS`, now independent of `ENABLE_OMOS` and the
+  presence of the bundled-skills source on the omos variant).
+- `entrypoint-user.sh`: both `oh-my-opencode-slim install` calls now pass
+  `--skills=no` — the installer manages only `oh-my-opencode-slim.json`.
+- One-time migration backs up (never deletes) any frozen real copies in
+  `~/.config/opencode/skills/` to `<name>.bak.<epoch>` (marker:
+  `~/.config/opencode/.omos-skills-migrated`).
+- `scripts/smoke-test.sh`: asserts the bundled-skills source path exists on the
+  omos variant (catches an upstream package restructure).
+- Docs: new `docs/omos-skills.md`; updated `README.md`, `AGENTS.md`,
+  `.env.example` (`OMOS_SKILLS` semantics).
+
+**Editing `entrypoint-user.sh` advances the base hash** — this release carries a
+full base rebuild and a `base-latest` re-promote.
+
+## v2.0.0 — 2026-06-13
+
+**Major release: pi is fully removed from opencode-devbox** (deprecated in
+v1.17.2), and the user-writable global npm prefix is relocated off the
+pi-specific `~/.pi` path. Also bumps opencode. The major version signals the
+two breaking changes below.
+
+### Bumped: opencode-ai 1.17.2 → 1.17.4
+
+`OPENCODE_VERSION` ARG in `Dockerfile.variant`.
+
+### Removed: all pi support
+
+pi ships as its own self-contained image,
+[`joakimp/pi-devbox`](https://gitea.jordbo.se/joakimp/pi-devbox). Everything
+pi-related is now gone from this repo:
+
+- The `INSTALL_PI` build arg and all `PI_*` args (`PI_VERSION`,
+  `PI_TOOLKIT_REF`, `PI_EXTENSIONS_REF`, `PI_FORK_*`, `PI_OBSMEM_*`).
+- The `with-pi`, `omos-with-pi`, and `pi-only` build variants. **Only `base`
+  and `omos` remain** — four published tags per release (`vX.Y.Z`, `latest`,
+  `vX.Y.Z-omos`, `latest-omos`).
+- The `base-pi-only[-vX.Y.Z]` tag that this repo's CI published into the
+  `joakimp/pi-devbox` repo. The publisher job is deleted; the orphaned
+  `base-pi-only*` tags on the pi-devbox Hub repo can now be purged (see
+  `docs/CLEANUP-v2.0.0.md`).
+- All `~/.pi` entrypoint wiring (pi-toolkit / pi-extensions deploy,
+  settings.json bootstrap, mempalace pi-bridge symlink, pi-fork / pi-obsmem
+  registration), the `~/.pi` volume-ownership entry, and the three pi
+  smoke / validate / build-variant CI jobs.
+
+**Migration:** pull `joakimp/pi-devbox:latest` instead of any `*-with-pi` /
+`pi-only` opencode-devbox tag. opencode-only users are unaffected by the pi
+removal itself.
+
+### Breaking: global npm prefix relocated `~/.pi/npm-global` → `~/.config/opencode/npm-global`
+
+`NPM_CONFIG_PREFIX` (and the matching `PATH` entry) moved off the
+pi-specific path. The new location lives on the `devbox-opencode-config`
+named volume, which — unlike the old `devbox-pi-config` — is a **persistent
+named volume in both `docker-compose.yml` and `docker-compose.shared.yml`**,
+so runtime `npm install -g` still survives container recreate and image
+rebuild.
+
+**Impact on upgraders:** any tool you previously `npm install -g`'d landed in
+`~/.pi/npm-global` and will **drop off `PATH`** under v2.0.0.
+
+**Mitigation (automatic):** `entrypoint-user.sh` carries a one-time
+migration shim. On first start, if `~/.pi/npm-global` exists and the marker
+`~/.config/opencode/npm-global/.migrated-from-dot-pi` is absent, it copies
+the old `lib/`/`bin/`/`share/` contents into the new prefix (never
+overwriting freshly-installed packages) and writes the marker. For the shim
+to see your old packages, the legacy `devbox-pi-config` volume must still be
+mounted at `~/.pi` for that first start — the shipped `docker-compose.yml`
+leaves it commented with instructions; uncomment the mount for one start if
+you had global npm tools to migrate, then remove it. Fresh installs carry no
+`~/.pi` dead weight. If you don't need the old packages, ignore all of this
+and re-`npm install -g` anything you want.
+
+### Docs / CI
+
+- `DOCKER_HUB.md` now shows the baked opencode version via a
+  `{{OPENCODE_VERSION}}` placeholder substituted by CI at publish time
+  (mirrors pi-devbox's `{{PI_VERSION}}` pattern) — the Hub page can no longer
+  drift from the image.
+- README, AGENTS.md, `.gitea/README.md`, `.env.example`, and the
+  `docs/manual-host-publish.*` runbook updated to the two-variant reality
+  (variant tables, CI job lists, the npm-prefix gotcha, ASCII pipeline
+  diagram, tag counts).
+- CI: removed the three pi smoke jobs, three pi build-variant jobs, the
+  `pi-only` publish-to-pi-devbox job, and the pi/fork/obsmem resolution in
+  `resolve-versions` (now resolves omos only). `validate.yml` drops its
+  three pi validate jobs.
+
+## v1.17.2 — 2026-06-10
+
+First container build on **opencode-ai `1.17.2`** (from `1.16.2`). This
+release also **deprecates all pi support** ahead of its removal in v2.0.0,
+and hardens the mempalace install.
+
+### Bumped: opencode-ai 1.16.2 → 1.17.2
+
+`OPENCODE_VERSION` ARG in `Dockerfile.variant`. Bare `v1.17.2` tag per the
+`v{opencode_version}` scheme.
+
+### Deprecated: pi support (removed in v2.0.0)
+
+pi has been decoupled into its own self-contained image,
+[`joakimp/pi-devbox`](https://gitea.jordbo.se/joakimp/pi-devbox) (v1.0.0+,
+which no longer FROMs `base-pi-only`). The pi paths in opencode-devbox are
+now dead weight and are **deprecated as of v1.17.2, scheduled for removal in
+v2.0.0**:
+
+- The `INSTALL_PI` build arg and all `PI_*` args.
+- The `with-pi`, `omos-with-pi`, and `pi-only` build variants.
+- The `base-pi-only[-vX.Y.Z]` tag published (to the `joakimp/pi-devbox`
+  repo) from this repo's CI.
+- All `~/.pi`-related entrypoint wiring (pi-toolkit / pi-extensions deploy,
+  settings.json bootstrap, mempalace pi-bridge symlink, pi-fork / pi-obsmem
+  registration).
+
+What this release does:
+
+- Building with `INSTALL_PI=true` now prints a **build-time deprecation
+  warning** to stderr.
+- README, DOCKER_HUB.md, and AGENTS.md mark the pi variants deprecated and
+  point to `joakimp/pi-devbox`.
+- The full removal plan is documented in `docs/CLEANUP-v2.0.0.md`.
+
+**Migration:** pull `joakimp/pi-devbox:latest` directly instead of any
+`*-with-pi` / `pi-only` opencode-devbox tag. opencode-only users are
+unaffected by this release.
+
+#### ⚠ Heads-up for v2.0.0: global npm prefix relocation
+
+Today `NPM_CONFIG_PREFIX` points at the pi-specific `~/.pi/npm-global`
+(backed by the `devbox-pi-config` named volume), so `npm install -g` as the
+developer user persists across recreates. **v2.0.0 will move the prefix to a
+neutral opencode path** (e.g. `~/.config/opencode/npm-global`). Consequences
+for existing opencode users at the v2.0.0 upgrade:
+
+- Previously global-installed npm tools remain on disk in the old volume but
+  **drop off `PATH`** until migrated.
+- The new prefix path is not currently a named volume, so persistence needs a
+  compose/volume update.
+
+v2.0.0 will ship a **one-time migration shim** (copies old prefix contents to
+the new path on first run) and an updated volume mapping. This notice is the
+one-release-cycle advance warning.
+
+### Hardened: mempalace install pinned + diary_write schema workaround
+
+- `MEMPALACE_VERSION` is now an explicit ARG (default `3.4.0`) in
+  `Dockerfile.base`; the install uses `mempalace==${MEMPALACE_VERSION}`
+  instead of an unpinned `uv tool install mempalace`. An unpinned install is
+  what silently swept in the broken `diary_write` schema. Mirrors pi-devbox.
+- Added an idempotent, self-deactivating post-install patch that strips the
+  **top-level `anyOf`** from `mempalace_diary_write`'s `input_schema`.
+  Mempalace 3.3.x/3.4.0 advertise `anyOf:[{required:[entry]},{required:[content]}]`
+  at the schema root, which the Anthropic tools API rejects outright
+  (`input_schema does not support oneOf, allOf, or anyOf at the top level`),
+  breaking pi/Claude tool registration at session start. The handler still
+  accepts `content` server-side. Upstream: MemPalace/mempalace#1728, PR #1735.
+  Remove once a fixed mempalace release is pinned.
+
+### Fixed: smoke-test pi-extensions readiness race (test-only, no image change)
+
+`scripts/smoke-test.sh`'s entrypoint-deploy wait loop gated only on
+`keybindings.json` (written by pi-toolkit, which runs *before* pi-extensions),
+so the `~/.pi/agent/extensions/*.ts ≥ 4` assertion could sample mid-deploy and
+see fewer than 4 files under parallel build load. Observed on v1.16.2 run 370:
+`smoke-with-pi` saw `<4` while `smoke-omos-with-pi` / `smoke-pi-only` (identical
+pi-extensions `357fcc6`) both saw 8, skipping `build-variant-with-pi`. The wait
+loop now blocks until the *last*-deployed artifact (the `mempalace.ts` bridge
+symlink) exists **and** the extension count has settled ≥ 4 (up to 45s).
+
+## v1.16.2 — 2026-06-08
+
+First container build on the `opencode-ai@1.16.x` minor release (rolls up
+`1.16.0` → `1.16.1` → `1.16.2`, all published 2026-06-05). Also picks up
+**pi `0.78.1` → `0.79.0`** (resolved fresh by CI's `resolve-versions` job) in
+the `with-pi`, `omos-with-pi`, and `pi-only` variants.
+
+### Bumped: opencode-ai 1.15.13 → 1.16.2
+
+`OPENCODE_VERSION` ARG in `Dockerfile.variant`. Highlights from the upstream
+release (full notes: <https://github.com/anomalyco/opencode/releases>):
+
+- **1.16.0** — ~38% faster startup (@StarpTech); managed workspace cloning that
+  keeps dirty/untracked files; move sessions between workspaces/directories;
+  proper OpenAI-via-Bedrock support; skill discovery + file-based agent loading;
+  `run --replay` for interactive session replay. Plus TUI/desktop polish and
+  numerous bugfixes (shell cancellation races, Windows path normalization, ACP
+  cancel/abort).
+- **1.16.1** — internal/no user-visible notes (empty release body).
+- **1.16.2** — reasoning summaries only run on supporting providers (avoids
+  GPT-5 request failures); edit operations refuse loose matches that could
+  overwrite the wrong code; Bedrock hang-before-first-token fix; diff-viewer
+  hunk navigation; subagents can be backgrounded; Snowflake Cortex provider.
+
+### Picks up: pi 0.78.1 → 0.79.0
+
+Resolved at build time for the pi-bearing variants. Headlines: project-trust
+prompting for project-local settings/resources/instructions/packages (with
+`--approve`/`--no-approve` and a `project_trust` extension event), cache-hit
+rate in the interactive footer, richer SDK/RPC extension surfaces, plus a
+stack of TUI and provider fixes. Full notes ship in the npm tarball's
+`CHANGELOG.md`.
+
+### Smoke size thresholds bumped +150 MB (preemptive)
+
+Ahead of the combined minor opencode + pi bump, all opencode-bearing variant
+thresholds in `scripts/smoke-test.sh` were raised: `base` 2600 → 2750,
+`omos` 3300 → 3450, `with-pi` 2900 → 3050, `omos-with-pi` 3900 → 4050; and
+`pi-only` 2750 → 2850 (+100, pi-only carries only the pi bump). Both `base`
+(last 2506 MB) and `omos` (last 3206 MB) were on ~94 MB headroom, and a minor
+opencode bump has tripped these ceilings before (v1.15.0 omos, v1.15.4
+omos-with-pi), causing a partial publish + letter-suffix recovery cycle.
+Restoring ~250 MB headroom avoids that. CI's smoke size print records actual
+landed sizes — tighten later if they come in well under.
+
+## v1.15.13e — 2026-06-04
+
+Letter-suffix rebuild on opencode `1.15.13` (version unchanged). Picks up
+**pi `0.78.1`** (resolved fresh by CI's `resolve-versions` job) plus the LAN-jump
+key-persistence work, an entrypoint ownership fix for the new `devbox-ssh-local`
+volume, a CI smoke false-negative fix, and documentation. Touches `entrypoint.sh`
+and `setup-lan-access.sh` (both in the base hash), so `base-latest` /
+`base-pi-only` advance and the fixes propagate to `pi-devbox`.
+
+### Docs: per-host `ControlPath` overrides break `pi --ssh` (read-only `~/.ssh`)
+
+Documented a gotcha in the README "Reaching your LAN" section: the bind-mounted
+`~/.ssh/config` is read before the baked `Host *` default, and SSH uses the
+first `ControlPath` it sees. A per-host block that sets `ControlPath` under
+`~/.ssh/` (a common CGNAT-multiplexing pattern, e.g. `~/.ssh/cm/%r@%h:%p`) wins
+but then fails inside the container because `~/.ssh` is mounted read-only — the
+master socket can't bind. This silently breaks `pi --ssh <host>`: the SSH layer
+fails and pi falls back to running its tools locally in the container. Fix is
+host-side — drop the per-host `ControlPath` or repoint it at the writable
+`/tmp/sshcm/%r@%h:%p` (works on both host and container, preserves multiplexing).
+No image change; documentation only.
+
+### Fixed: validate.yml false-negative on fork/recall registration checks
+
+The push-to-main `validate.yml` builds variants FROM the published `base-latest`
+image, which lags the entrypoint in the current commit until a release tag
+rebuilds the base. The fork/recall *registration* smoke checks depend on the
+base entrypoint running `pi install /opt/<pkg>`, so a stale `base-latest` reded
+those runs with a false negative even when the variant layer was correct.
+`smoke-test.sh` now gates the two registration assertions behind
+`STRICT_REGISTRATION` (warn-only when unset). `validate.yml` leaves it unset;
+the release pipeline (`docker-publish-split.yml`), which builds the base fresh
+in the same run, sets `STRICT_REGISTRATION=1` on the pi-bearing smoke jobs to
+enforce them. The build-time `/opt` + `node_modules` checks stay hard in both
+paths.
+
+### Added: persist the LAN-jump key + one-line authorize hint (authorize once per machine)
+
+The jump keypair (`~/.ssh-local/devbox_jump_ed25519`) was stored on the
+container's ephemeral overlay, so `docker compose up --force-recreate` (every
+image update) regenerated it — forcing you to re-authorize the new key on the
+host each time. The compose files now persist `~/.ssh-local` via a named volume
+(`devbox-ssh-local`), matching the pattern already used for `.pi`, shell
+history, etc. The key is generated **once** and reused across updates, so you
+authorize it on the host **once per machine**.
+
+`setup-lan-access.sh` now also prints a ready-to-paste authorize line whenever
+it generates a **new** key (not just when `HOST_SSH_USER` is unset), e.g.
+`echo 'ssh-ed25519 …' >> ~/.ssh/authorized_keys` — no helper file to locate, no
+workspace path to guess. It stays silent once the key is persisted.
+
+### Fixed: chown the `devbox-ssh-local` volume so the jump key can be generated
+
+The previous change persisted `~/.ssh-local` via a named volume, but the
+entrypoint's volume-ownership loop was never updated to include it. Docker
+creates named volumes as `root:root`, so on a fresh volume `~/.ssh-local`
+stayed root-owned while `setup-lan-access.sh` runs as `developer` — both its
+`mkdir cm` and `ssh-keygen` failed silently (`|| true` / `|| exit 0`), leaving
+**no jump key and no config**, breaking LAN access on the first recreate after
+the persistence change. `entrypoint.sh` now chowns `~/.ssh-local` to the
+developer user alongside the other named-volume mount points.
+
+### Docs: document the optional `~/.config/devbox-shell` mount in the compose template
+
+`docker-compose.yml` now carries a commented-out `~/.config/devbox-shell` bind
+mount with an explanatory note. It's the recommended home for host-owned shell
+config: the image's `~/.bash_aliases` sources `~/.config/devbox-shell/bash_aliases`
+if present, and `setup-lan-access.sh` reads `~/.config/devbox-shell/ssh-lan.conf`
+for named-peer `ProxyJump host` overrides. A directory mount is preferred over
+the single-file `~/.bash_aliases` mount because it survives editors' atomic-save.
+Template comment only; no behavior change.
+
+## v1.15.13d — 2026-06-04
+
+LAN-access fixes + ergonomics. Letter-suffix rebuild on opencode `1.15.13`
+(version unchanged). Touches `setup-lan-access.sh`, which is in the base hash,
+so `base-latest` / `base-pi-only` advance and the fix propagates to `pi-devbox`.
+
+### Fixed: LAN-access `Include` was scoped to the `host`/`mac` block (named peers ignored)
+
+The generated `~/.ssh-local/config` placed `Include ~/.ssh/config` *inside* the
+`Host host mac` block. Because SSH scopes an `Include` to the enclosing
+`Host`/`Match` block, the user's `~/.ssh/config` was only consulted when
+targeting `host`/`mac` — so `dssh pve` / `dssh <peer>` by name silently fell
+back to SSH defaults (wrong user, unresolved hostname) and never applied the
+peer's settings or any `ProxyJump`. Fixed by emitting a bare `Host *` scope
+reset before every `Include`.
+
+### Fixed: read-only `~/.ssh/cm` ControlPath broke multiplexed hosts
+
+The bind-mounted `~/.ssh/config` commonly sets `ControlPath ~/.ssh/cm/...`
+(CGNAT flow-cap multiplexing), but `~/.ssh` is read-only in the container, so
+every `ControlMaster`-enabled host (e.g. `pmx-jh`, `proxmox*`, `synlig`) failed
+with `cannot bind to path … Read-only file system`. The generated config now
+sets `ControlPath ~/.ssh-local/cm/%r@%h:%p` in the top `Host *` block
+(first-value-wins) so master sockets land in the writable sidecar.
+
+### Added: host-owned `ssh-lan.conf` for named-peer jump overrides
+
+When the host bind-mounts `~/.config/devbox-shell/ssh-lan.conf`, the generated
+config now Includes it *before* `~/.ssh/config`. Put `ProxyJump host` overrides
+there (first-value-wins inherits HostName/User/IdentityFile from `~/.ssh/config`)
+instead of editing the shared `~/.ssh/config` — which would break the host's own
+direct access to those peers and is read-only from the container anyway. New
+[`ssh-lan.conf.example`](ssh-lan.conf.example).
+
+### Added: `DEVBOX_LAN_AUTOJUMP_PRIVATE=1` opt-in RFC1918 auto-jump
+
+Emits a catch-all that ProxyJumps any private (RFC1918) IP through the host, so
+bare `dssh user@<ip>` reaches whatever LAN the (roaming) host is currently on,
+without naming peers. Matches the typed address (not the resolved HostName), so
+named hosts carrying their own ProxyJump are unaffected; public IPs stay direct.
+
+All three land in `rootfs/usr/local/lib/opencode-devbox/setup-lan-access.sh`,
+which is counted in the base hash → advances `base-latest` and propagates to
+`pi-devbox` (built `FROM` the base).
 
 ## v1.15.13c — 2026-06-03
 

DOCKER_HUB.md

@@ -2,23 +2,24 @@
 
 Portable AI developer environment for [opencode](https://opencode.ai). Debian-based, with git, SSH, Node.js, AWS CLI v2, and common dev tools pre-installed.
 
+> **Current `:latest` ships opencode `{{OPENCODE_VERSION}}`** (the baked version is asserted by smoke tests, so this page never drifts from the image).
+
 Designed for teams who want a reproducible coding-agent setup that runs the same on every laptop and CI runner — without forcing each developer to install Bun, Node, AWS CLI, mempalace, or maintain shell config drift across machines.
 
 ## Image Variants
 
 | Tag | Description |
 |---|---|
-| `latest` / `vX.Y.Z` | Base image — opencode, Node.js, AWS CLI, dev tools |
+| `latest` / `vX.Y.Z` | Base image — opencode `{{OPENCODE_VERSION}}`, Node.js, AWS CLI, dev tools |
 | `latest-omos` / `vX.Y.Z-omos` | Base + [oh-my-opencode-slim](https://github.com/alvinunreal/oh-my-opencode-slim) multi-agent orchestration and Bun |
-| `latest-with-pi` / `vX.Y.Z-with-pi` | Base + [pi](https://github.com/earendil-works/pi) as alternative/complementary harness (shares the mempalace install with opencode) |
-| `latest-omos-with-pi` / `vX.Y.Z-omos-with-pi` | OMOS + pi together |
 
 All variants support `linux/amd64` and `linux/arm64`.
 
-> A fifth, pi-without-opencode build is produced from the same `Dockerfile.variant`
-> (`INSTALL_OPENCODE=false`) but is **not** published under this repo — it ships as
-> the separate [`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox)
-> image so an "opencode-devbox" tag never lacks opencode.
+> **Looking for pi?** As of v2.0.0 the pi coding-agent is no longer bundled in
+> opencode-devbox. It ships as the dedicated
+> [`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox) image, which
+> shares the same mempalace memory layer. See
+> <https://gitea.jordbo.se/joakimp/pi-devbox>.
 
 ## Quick Start
 
@@ -33,7 +34,7 @@ curl -fsSL https://gitea.jordbo.se/joakimp/opencode-devbox/raw/branch/main/.env.
 docker compose run --rm devbox
 ```
 
-This drops you straight into opencode with your project mounted at `/workspace`. Use `bash` as the command (e.g. `docker compose run --rm devbox bash`) to land in a shell first — useful for `aws sso login`, `pi` (on `*-with-pi` variants), or multi-harness workflows.
+This mounts your project at `/workspace`. With no command (as above) the image's default `CMD` (`bash -l`) drops you into a login shell — run `opencode` to start the harness, or `aws sso login` first, etc. To start opencode directly, pass it as the command: `docker compose run --rm devbox opencode`.
 
 **One-shot run, no persistence:**
 
@@ -53,11 +54,10 @@ Full setup guide — authentication for each provider (Anthropic, OpenAI, Bedroc
 ## What's Inside
 
 - **[opencode](https://opencode.ai)** — primary coding-agent harness. Multi-provider (Anthropic, OpenAI, Bedrock, Google, Groq, etc.).
-- **[pi](https://github.com/earendil-works/pi)** *(in `*-with-pi` variants)* — lightweight TUI coding-agent that coexists with opencode and shares the same mempalace install. Includes the `mcp-loader` extension so any local-stdio or remote streamable-HTTP MCP server (searxng, gitea, context7, …) can be added by editing `~/.pi/agent/settings.json`.
-- **[mempalace](https://github.com/MemPalace/mempalace)** — persistent AI memory layer (ChromaDB + SQLite). Wing/diary/knowledge-graph entries are mutually visible to opencode and pi.
+- **[mempalace](https://github.com/MemPalace/mempalace)** — persistent AI memory layer (ChromaDB + SQLite). Wing/diary/knowledge-graph entries are shareable with the sibling [`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox) image when both point at the same palace.
 - **[oh-my-opencode-slim](https://github.com/alvinunreal/oh-my-opencode-slim)** *(in `*-omos` variants)* — multi-agent orchestration on top of opencode (council, fallback chains, named agents).
 - **AWS CLI v2** with SSO support, **Node.js LTS**, **Bun** (OMOS variants), **uv** (Python), **gosu** for clean UID/GID adjustment to match your host workspace.
-- **MCP wrappers** for mempalace pre-installed and pre-wired to both harnesses.
+- **MCP wrappers** for mempalace pre-installed and pre-wired to opencode.
 
 ## Authentication
 
@@ -75,8 +75,7 @@ https://gitea.jordbo.se/joakimp/opencode-devbox#aws-bedrock-authentication
 
 | Volume | Mount | Survives |
 |---|---|---|
-| `devbox-opencode-config` | `~/.config/opencode` | container recreate, image rebuild |
-| `devbox-pi-config` | `~/.pi` | container recreate, image rebuild — incl. user-installed pi packages via `pi install` (`NPM_CONFIG_PREFIX` points into the volume) |
+| `devbox-opencode-config` | `~/.config/opencode` | container recreate, image rebuild — incl. user-installed npm globals via `npm install -g` (`NPM_CONFIG_PREFIX` points into the volume) |
 | `devbox-palace` (uncomment) | `~/.mempalace` | container recreate, image rebuild — palace data is precious, treat as primary storage |
 | `devbox-chroma-cache` | `~/.cache/chroma` | container recreate (model cache, disposable — re-downloads in seconds) |
 
@@ -93,7 +92,7 @@ Full persistence reference, including multi-user (`SIGNUM`) isolation and host b
 
 ## Sibling images
 
-- **[`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox)** — pi-only image built on top of this image's base layer. Smaller (~700 MB) and version-tracks the [pi npm package](https://www.npmjs.com/package/@earendil-works/pi-coding-agent) directly. Use this if you want pi without opencode. Source: <https://gitea.jordbo.se/joakimp/pi-devbox>
+- **[`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox)** — the pi coding-agent in its own self-contained image, built on a shared Debian base. Version-tracks the [pi npm package](https://www.npmjs.com/package/@earendil-works/pi-coding-agent) directly and can share this image's mempalace palace. Use it if you want pi instead of (or alongside) opencode. Source: <https://gitea.jordbo.se/joakimp/pi-devbox>
 
 ## License
 
@@ -101,4 +100,4 @@ MIT. See <https://gitea.jordbo.se/joakimp/opencode-devbox/src/branch/main/LICENS
 
 ---
 
-> This description is generated by `scripts/generate-dockerhub-md.py` from a hand-maintained template. Edit the template (not this file) and regenerate.
+> This description is generated by `scripts/generate-dockerhub-md.py` from a hand-maintained template. Edit the template (not this file) and regenerate. The `{{OPENCODE_VERSION}}` placeholder is filled by CI at publish time.

Dockerfile.base

@@ -1,22 +1,21 @@
 # opencode-devbox — base image (variant-independent layers)
 #
 # This Dockerfile produces an image tagged base-<hash>, used as the parent
-# for all four published variants (base, omos, with-pi, omos-with-pi).
-# It contains everything that does not depend on variant-specific
-# build-args (INSTALL_OPENCODE, INSTALL_OMOS, INSTALL_PI). The variant
-# Dockerfile (Dockerfile.variant) FROMs the base and adds only those
-# deltas.
+# for all published variants (base, omos). It contains everything that
+# does not depend on variant-specific build-args (INSTALL_OPENCODE,
+# INSTALL_OMOS). The variant Dockerfile (Dockerfile.variant) FROMs the
+# base and adds only those deltas.
 #
 # The base is rebuilt only when this file or anything it COPYs in
 # changes (rootfs/, entrypoint*.sh). Version bumps to OPENCODE_VERSION,
-# OMOS_VERSION, PI_VERSION, etc. do NOT trigger a base rebuild.
+# OMOS_VERSION, etc. do NOT trigger a base rebuild.
 #
 # To force a base rebuild for fresh apt packages without other code
 # changes, bump the BASE_REBUILD_DATE comment below. The hash is
 # content-addressed over this file, so any byte change invalidates the
 # cache. Recommended cadence: once per release for security updates.
 #
-# BASE_REBUILD_DATE: 2026-05-14 (v1.14.50b — fresh apt + first promote-base-latest)
+# BASE_REBUILD_DATE: 2026-07-13 (v2.7.0 — typst PDF engine + terminal terminfo (ncurses-term/kitty/xterm-ghostty) + nvim true-colour; pi-devbox parity)
 #
 # See the project README's "Build pipeline" section for the rationale.
 
@@ -36,6 +35,11 @@ ENV DEBIAN_FRONTEND=noninteractive
 # apt-get upgrade picks up any security/CVE fixes published between
 # debian:trixie-slim base-image rebuilds. Paired with the index update
 # and the install in the same layer so we don't bloat image history.
+# `nano` is included as a small, non-modal terminal editor for users who
+# don't want vi-style modal editing — a companion to nvim and the `micro`
+# binary installed further down. ~2.8 MB; its deps (libc6, libncursesw6,
+# libtinfo6) are already pulled in by nvim/less/htop/tmux, so it adds no
+# extra packages. EDITOR stays nvim; opt in via `export EDITOR=nano`.
 RUN apt-get update && \
     apt-get upgrade -y --no-install-recommends && \
     apt-get install -y --no-install-recommends \
@@ -67,6 +71,12 @@ RUN apt-get update && \
     rsync \
     python3-pip \
     python3-venv \
+    nano \
+    pandoc \
+    xz-utils \
+    graphviz \
+    kitty-terminfo \
+    ncurses-term \
     && ln -s /usr/bin/fdfind /usr/local/bin/fd \
     && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
@@ -95,6 +105,15 @@ RUN apt-get update && \
 # the last session closes, so consecutive ssh calls in a workflow reuse
 # the same TCP flow. Companion entrypoint-user.sh creates /tmp/sshcm
 # (mode 700) on each container start.
+#
+# CAVEAT (and why dssh/dscp are handled elsewhere): a user per-host override
+# that points ControlPath BACK under the read-only ~/.ssh (e.g. the common
+# CGNAT idiom `ControlPath ~/.ssh/cm/%r@%h:%p`) re-introduces the
+# unwritable-socket failure for a plain `ssh <host>` — a system drop-in here
+# can never override a user's per-host value. For `ssh -F ~/.ssh-local/config`
+# (the dssh/dscp aliases), setup-lan-access.sh redirects ControlPath into the
+# writable ~/.ssh-local sidecar, so those paths are unaffected. See CHANGELOG
+# "Unreleased".
 RUN mkdir -p /etc/ssh/ssh_config.d && \
     printf '%s\n' \
       '# Devbox-baked default. See Dockerfile.base "SSH client defaults".' \
@@ -196,6 +215,33 @@ RUN ARCH=$(case "${TARGETARCH}" in amd64) echo "x86_64" ;; arm64) echo "arm64" ;
     ln -s /opt/nvim-linux-${ARCH}/bin/nvim /usr/local/bin/nvim && \
     nvim --version | head -1
 
+# micro — modern, non-modal terminal editor. Ships alongside nvim so users
+# who aren't comfortable with vi-style modal editing have a friendly option:
+# desktop-style keybindings (Ctrl+S save, Ctrl+Q quit, Ctrl+C/V/X, Ctrl+Z
+# undo), mouse support, and syntax highlighting out of the box. A single
+# static Go binary (~12 MB) installed from GitHub releases, exactly like
+# bat/eza/zoxide below. EDITOR stays nvim (see below); users opt in with
+# `export EDITOR=micro` or `git config --global core.editor micro`.
+#
+# NOTE: upstream moved zyedidia/micro -> micro-editor/micro. The old org URL
+# still 302s, but its /releases/latest redirect lands on ANOTHER /latest URL
+# (the org rename), so the tag-parsing idiom below would resolve "latest"
+# instead of a version. Use the canonical micro-editor/micro URL.
+# Arch asset naming differs from the others: amd64 -> linux64, arm64 ->
+# linux-arm64. The tarball extracts to micro-<version>/micro.
+ARG MICRO_VERSION=latest
+RUN ARCH=$(case "${TARGETARCH}" in amd64) echo "linux64" ;; arm64) echo "linux-arm64" ;; *) echo "linux64" ;; esac) && \
+    V="${MICRO_VERSION}" && \
+    if [ "$V" = "latest" ]; then \
+      V=$(curl -sI --retry 5 --retry-delay 5 --retry-all-errors "https://github.com/micro-editor/micro/releases/latest" | awk 'tolower($1)=="location:" { sub(/\r$/,"",$2); n=split($2,a,"/"); print a[n] }'); \
+    fi && \
+    V="${V#v}" && [ -n "$V" ] && \
+    echo "Installing micro ${V}" && \
+    curl -fsSL --retry 5 --retry-delay 5 --retry-all-errors "https://github.com/micro-editor/micro/releases/download/v${V}/micro-${V}-${ARCH}.tar.gz" | tar -xz -C /tmp && \
+    install /tmp/micro-${V}/micro /usr/local/bin/micro && \
+    rm -rf /tmp/micro-${V} && \
+    micro --version
+
 # bat — syntax-highlighted cat replacement
 ARG BAT_VERSION=latest
 RUN ARCH=$(case "${TARGETARCH}" in amd64) echo "x86_64" ;; arm64) echo "aarch64" ;; *) echo "x86_64" ;; esac) && \
@@ -254,25 +300,143 @@ RUN ARCH=$(case "${TARGETARCH}" in amd64) echo "x86_64" ;; arm64) echo "aarch64"
     rm -rf /tmp/uv-* && \
     uv --version
 
+# tealdeer — Rust port of tldr (community-maintained command examples)
+# Provides the `tldr` command; ~5 MB static binary, ~135 MB smaller than
+# the Node tldr global. Same UX as the Node version.
+ARG TEALDEER_VERSION=latest
+RUN ARCH=$(case "${TARGETARCH}" in amd64) echo "x86_64" ;; arm64) echo "aarch64" ;; *) echo "x86_64" ;; esac) && \
+    V="${TEALDEER_VERSION}" && \
+    if [ "$V" = "latest" ]; then \
+      V=$(curl -sI --retry 5 --retry-delay 5 --retry-all-errors "https://github.com/tealdeer-rs/tealdeer/releases/latest" | awk 'tolower($1)=="location:" { sub(/\r$/,"",$2); n=split($2,a,"/"); print a[n] }'); \
+    fi && \
+    V="${V#v}" && \
+    [ -n "$V" ] && \
+    echo "Installing tealdeer ${V}" && \
+    curl -fsSL --retry 5 --retry-delay 5 --retry-all-errors "https://github.com/tealdeer-rs/tealdeer/releases/download/v${V}/tealdeer-linux-${ARCH}-musl" -o /usr/local/bin/tldr && \
+    chmod +x /usr/local/bin/tldr && \
+    tldr --version
+
+# ── typst — lightweight PDF engine for pandoc (Markdown→PDF) ─────────
+# pandoc (apt-installed above) is only a front-end; rendering PDF needs a
+# back-end engine. Rather than a ~600 MB TeX Live install, we ship typst:
+# a single ~30 MB static Rust binary with no LaTeX dependency, used via
+# `pandoc --pdf-engine=typst`. A fuller TeX Live remains the higher-fidelity
+# fallback for anyone who needs LaTeX-exact output (install on demand).
+# Ported from pi-devbox (its v1.4.0 + v1.5.0 font-default fix).
+#
+# Follows the `latest` GitHub-release convention (like tealdeer/uv/bat).
+# typst ships a `.tar.xz` asset (hence xz-utils in the apt layer above)
+# that extracts to typst-<arch>-unknown-linux-musl/typst. Pin a specific
+# tag with --build-arg TYPST_VERSION=vX.Y.Z.
+#
+# We also patch pandoc's bundled typst template
+# (/usr/share/pandoc/data/templates/template.typst): its conf() defaults the
+# document font to an empty tuple (`font: ()`), so a naked
+# `pandoc --pdf-engine=typst` fails with "font fallback list must not be empty"
+# unless the caller passes `-V mainfont=...`. We default it to Libertinus Serif
+# (typst's own bundled default font) so PDF export works out-of-the-box.
+ARG TYPST_VERSION=latest
+RUN ARCH=$(case "${TARGETARCH}" in amd64) echo "x86_64" ;; arm64) echo "aarch64" ;; *) echo "x86_64" ;; esac) && \
+    V="${TYPST_VERSION}" && \
+    if [ "$V" = "latest" ]; then \
+      V=$(curl -sI --retry 5 --retry-delay 5 --retry-all-errors "https://github.com/typst/typst/releases/latest" | awk 'tolower($1)=="location:" { sub(/\r$/,"",$2); n=split($2,a,"/"); print a[n] }'); \
+    fi && \
+    V="${V#v}" && [ -n "$V" ] && \
+    echo "Installing typst ${V}" && \
+    curl -fsSL --retry 5 --retry-delay 5 --retry-all-errors "https://github.com/typst/typst/releases/download/v${V}/typst-${ARCH}-unknown-linux-musl.tar.xz" | tar -xJ -C /tmp && \
+    install /tmp/typst-${ARCH}-unknown-linux-musl/typst /usr/local/bin/typst && \
+    rm -rf /tmp/typst-${ARCH}-unknown-linux-musl && \
+    typst --version && \
+    sed -i 's/^  font: (),$/  font: ("Libertinus Serif",),/' /usr/share/pandoc/data/templates/template.typst && \
+    grep -q 'font: ("Libertinus Serif",),' /usr/share/pandoc/data/templates/template.typst
+
+# ── yq (mikefarah) — YAML processor, jq's companion for YAML ─────────
+# Installed as the mikefarah Go binary — NOT Debian's `yq` apt package, which
+# is the unrelated Python kislyuk/yq (a jq wrapper with different syntax and a
+# different version line, 3.x). THIRD_PARTY.md already credited "yq
+# (mikefarah)" while the image actually shipped the Python one, so this closes
+# a documented-vs-shipped mismatch as well as bringing parity with pi-devbox
+# (its v1.2.3). The cloud-init repo's deploy.sh/provision.sh require mikefarah
+# v4 syntax. Follows the repo's `latest` convention (like tealdeer/uv/typst);
+# the smoke test pins the contract to major v4, so both a regression to the
+# Python package and a surprise future yq v5 fail CI loudly instead of
+# silently breaking those scripts. Pin a tag with --build-arg YQ_VERSION=vX.Y.Z.
+ARG YQ_VERSION=latest
+RUN ARCH=$(case "${TARGETARCH}" in amd64) echo "amd64" ;; arm64) echo "arm64" ;; *) echo "amd64" ;; esac) && \
+    V="${YQ_VERSION}" && \
+    if [ "$V" = "latest" ]; then \
+      V=$(curl -sI --retry 5 --retry-delay 5 --retry-all-errors "https://github.com/mikefarah/yq/releases/latest" | awk 'tolower($1)=="location:" { sub(/\r$/,"",$2); n=split($2,a,"/"); print a[n] }'); \
+    fi && \
+    [ -n "$V" ] && \
+    echo "Installing mikefarah yq ${V}" && \
+    curl -fsSL --retry 5 --retry-delay 5 --retry-all-errors "https://github.com/mikefarah/yq/releases/download/${V}/yq_linux_${ARCH}" -o /usr/local/bin/yq && \
+    chmod +x /usr/local/bin/yq && \
+    yq --version
+
 # ── MemPalace — local-first AI memory system ─────────────────────────
 # Provides semantic search over conversation history via 29 MCP tools.
 # Always installed in the base (variant-independent). Set
 # INSTALL_MEMPALACE=false at base-build time to shave ~300 MB.
 ARG INSTALL_MEMPALACE=true
+# Pin mempalace explicitly (mirrors pi-devbox). An unpinned
+# `uv tool install mempalace` is what silently swept in the broken
+# diary_write top-level-anyOf schema (3.3.x/3.4.0) that breaks the
+# Anthropic tools API; pinning makes every bump a deliberate, reviewable
+# diff. Bump this in lockstep with pi-devbox's MEMPALACE_VERSION.
+#
+# 3.5.0 (2026-06) ships the upstream fix for that top-level-anyOf schema
+# (issue #1728 / PR #1717, merged 2026-06-14): diary_write now advertises
+# `"required": ["agent_name"]` with entry/content enforced at dispatch, which
+# the Anthropic tools API accepts — so the perl mcp_server.py workaround that
+# used to live below is gone. (pi-devbox dropped it in its v1.2.2.)
+#
+# 3.6.0 (2026-07-17) is an additive/reliability release — secure `mempalace
+# serve` remote mode, optional Milvus backend, atomic KG supersede(),
+# conversation chronology, mining exclusions, plus recovery/locking fixes.
+# Reviewed for MCP tool-schema changes before bumping: there are NONE, and
+# nothing touches diary_write — so the 3.3.x/3.4.0 regression class does not
+# recur. Two fixes are directly relevant to how this image uses mempalace:
+# read-only mode now covers checkpoint + delete_by_source in _MUTATING_TOOLS
+# (#1930), and agent attribution is preserved in mempalace_checkpoint
+# (#2023/#2034).
+ARG MEMPALACE_VERSION=3.6.0
 ENV UV_TOOL_DIR=/opt/uv-tools
 ENV UV_TOOL_BIN_DIR=/usr/local/bin
 RUN if [ "${INSTALL_MEMPALACE}" = "true" ]; then \
       mkdir -p /opt/uv-tools && \
-      uv tool install --no-cache mempalace && \
+      uv tool install --no-cache "mempalace==${MEMPALACE_VERSION}" && \
       /opt/uv-tools/mempalace/bin/python -c "import mempalace; print('mempalace', mempalace.__version__ if hasattr(mempalace, '__version__') else 'installed')" ; \
     fi
 
+# (The mempalace diary_write top-level-anyOf workaround that patched
+# mcp_server.py here was removed when MEMPALACE_VERSION moved to 3.5.0 —
+# fixed upstream via issue #1728 / PR #1717 (merged 2026-06-14). Mirrors
+# pi-devbox v1.2.2. See CHANGELOG.md.)
+
 # ── mempalace-toolkit — bash wrappers for session/docs mining ────────
 ARG INSTALL_MEMPALACE_TOOLKIT=true
 ARG MEMPALACE_TOOLKIT_REF=main
+# MEMPALACE_TOOLKIT_REPO is overridable so a relocated/forked build can repoint
+# the clone without editing this Dockerfile (matches the *_REPO pattern used by
+# other companions). Defaults to the canonical gitea origin; the default CI
+# build is byte-identical.
+ARG MEMPALACE_TOOLKIT_REPO=https://gitea.jordbo.se/joakimp/mempalace-toolkit.git
+# MEMPALACE_TOOLKIT_REF accepts EITHER a branch name OR a commit SHA. CI
+# resolves it to a SHA (resolve-versions job) and folds that SHA into the
+# base-decide hash so the base rebuilds when the toolkit moves. `git clone
+# --branch <40-char-SHA>` fails ("Remote branch not found"), so use
+# `git fetch <ref> + checkout FETCH_HEAD`, which works for name and SHA.
 RUN if [ "${INSTALL_MEMPALACE}" = "true" ] && [ "${INSTALL_MEMPALACE_TOOLKIT}" = "true" ]; then \
-      git clone --depth 1 --branch "${MEMPALACE_TOOLKIT_REF}" \
-        https://gitea.jordbo.se/joakimp/mempalace-toolkit.git /opt/mempalace-toolkit && \
+      rm -rf /opt/mempalace-toolkit && mkdir -p /opt/mempalace-toolkit && \
+      git -C /opt/mempalace-toolkit init -q && \
+      git -C /opt/mempalace-toolkit remote add origin "${MEMPALACE_TOOLKIT_REPO}" && \
+      ok=0; for i in 1 2 3 4 5; do \
+        if git -C /opt/mempalace-toolkit fetch --depth 1 origin "${MEMPALACE_TOOLKIT_REF}" && \
+           git -C /opt/mempalace-toolkit checkout -q FETCH_HEAD; then ok=1; break; fi; \
+        echo "git fetch mempalace-toolkit@${MEMPALACE_TOOLKIT_REF} failed (attempt $i/5), retrying in $((i*5))s..."; \
+        sleep $((i*5)); \
+      done; \
+      [ "$ok" = "1" ] && \
       ln -sf /opt/mempalace-toolkit/bin/mempalace-session /usr/local/bin/mempalace-session && \
       ln -sf /opt/mempalace-toolkit/bin/mempalace-docs    /usr/local/bin/mempalace-docs && \
       chmod +x /opt/mempalace-toolkit/bin/mempalace-session /opt/mempalace-toolkit/bin/mempalace-docs && \
@@ -307,7 +471,26 @@ ENV LANG=en_US.UTF-8
 ENV LANGUAGE=en_US:en
 ENV LC_ALL=en_US.UTF-8
 ENV EDITOR=nvim
+# Advertise 24-bit colour so colour-aware tools (Neovim's own auto-detect, bat,
+# delta, ...) use true colour instead of a 256-colour fallback. Completes the
+# true-colour story the terminfo + sysinit.vim layers below already start.
+# Safe for the modern terminals this devbox targets; override by exporting
+# `COLORTERM=` (empty) from a terminal that lacks true-colour support.
+ENV COLORTERM=truecolor
 ENV PATH="/home/developer/.local/bin:/home/developer/.cargo/bin:${PATH}"
+# Enable opencode's native background subagents. opencode gates this behind an
+# experimental flag (packages/opencode/src/tool/task.ts errors with
+# "Background subagents require OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=true"
+# when unset). oh-my-opencode-slim V2+ makes background orchestration its DEFAULT
+# workflow, so the omos variant is effectively broken without this. Baked here as
+# a base ENV (applies to both variants; harmless for plain opencode — it only
+# *enables* a capability). Overridable at runtime: -e OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=false.
+# REMOVAL TRIGGER: when opencode promotes background subagents out of experimental
+# (flag becomes default / renamed), drop this ENV. No upstream roadmap date as of
+# opencode 1.18.13 / omos 2.2.9 (2026-08). Re-verified against upstream source
+# at all three tags (1.17.20, 1.18.12, 1.18.13): packages/opencode/src/effect/runtime-flags.ts
+# is unchanged and still gates the flag behind OPENCODE_EXPERIMENTAL — trigger has NOT fired.
+ENV OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS=true
 
 # ── Node.js (required for opencode/pi/omos at variant build + MCP servers) ──
 ARG NODE_VERSION=22
@@ -315,6 +498,45 @@ RUN curl -fsSL --retry 5 --retry-delay 5 --retry-all-errors https://deb.nodesour
     apt-get install -y --no-install-recommends nodejs && \
     rm -rf /var/lib/apt/lists/*
 
+# ── agent-browser + Playwright Chromium — real-browser verification ──
+# Lets the agent drive an actual browser (open pages, click/fill/eval, snapshot
+# the DOM, screenshot) to VERIFY front-end work — live DOM, WebGL, layout,
+# popup positioning — instead of guessing. Ported from pi-devbox.
+#
+# We resolve the Chrome binary through a stable symlink (/usr/local/bin/
+# agent-chrome) exposed via AGENT_BROWSER_EXECUTABLE_PATH — the symlink
+# insulates the ENV from Playwright's per-version, per-ARCH browser directory
+# (`chrome-linux` on arm64, `chrome-linux64` on amd64 — Chrome-for-Testing), so
+# we `find` the `chrome` binary rather than hardcode the path; the headless-shell
+# binary is named `chrome-headless-shell`, so `-name chrome` skips it.
+#
+# `playwright install --with-deps chromium` also apt-installs Chromium's runtime
+# libs; verified to resolve correctly on Debian trixie (the t64 library renames
+# are handled by Playwright's dep list). The build runs as root, so the apt step
+# works. NPM_CONFIG_PREFIX=/usr keeps both CLIs on /usr so they survive the
+# ~/.config/opencode/npm-global volume mount (the same trick the variant uses
+# for opencode). After fetching, we DROP Playwright's `chromium_headless_shell-*`
+# build — agent-browser drives the full chrome, so the headless shell is dead
+# weight — and clean the apt/npm caches, trimming the layer to ~625 MB from
+# ~960 MB. This is the bulk of the base's size and the one real tradeoff of
+# shipping it to every variant; the smoke-test size thresholds were lifted in
+# lockstep (see scripts/smoke-test.sh).
+ARG AGENT_BROWSER_VERSION=latest
+ARG PLAYWRIGHT_VERSION=latest
+ENV PLAYWRIGHT_BROWSERS_PATH=/usr/local/share/ms-playwright
+RUN NPM_CONFIG_PREFIX=/usr npm install -g \
+      "agent-browser@${AGENT_BROWSER_VERSION}" \
+      "playwright@${PLAYWRIGHT_VERSION}" && \
+    playwright install --with-deps chromium && \
+    CHROME="$(find "${PLAYWRIGHT_BROWSERS_PATH}" -type f -name chrome -path '*/chromium-*/*' | head -n1)" && \
+    [ -n "$CHROME" ] && ln -sf "$CHROME" /usr/local/bin/agent-chrome && \
+    agent-browser --version && \
+    test -x "$(readlink -f /usr/local/bin/agent-chrome)" && \
+    rm -rf "${PLAYWRIGHT_BROWSERS_PATH}"/chromium_headless_shell-* && \
+    npm cache clean --force && \
+    rm -rf /var/lib/apt/lists/* /root/.npm /tmp/*
+ENV AGENT_BROWSER_EXECUTABLE_PATH=/usr/local/bin/agent-chrome
+
 # ── AWS CLI v2 (for SSO/Bedrock authentication) ─────────────────────
 RUN ARCH=$(case "${TARGETARCH}" in \
       amd64) echo "x86_64" ;; \
@@ -339,7 +561,7 @@ RUN groupadd --gid ${USER_GID} ${USER_NAME} && \
 # Create standard directories
 RUN mkdir -p /workspace \
     /home/${USER_NAME}/.config/opencode/skills \
-    /home/${USER_NAME}/.pi/agent/extensions \
+    /home/${USER_NAME}/.config/opencode/npm-global \
     /home/${USER_NAME}/.agents/skills \
     /home/${USER_NAME}/.local/share/opencode \
     /home/${USER_NAME}/.cache/bash \
@@ -359,31 +581,69 @@ print('chromadb embedding model warmed: all-MiniLM-L6-v2')" && \
       ls -lh /home/${USER_NAME}/.cache/chroma/onnx_models/all-MiniLM-L6-v2/ ; \
     fi
 
-# ── User-writable npm global prefix on the devbox-pi-config volume ──
+# ── User-writable npm global prefix on the devbox-opencode-config volume ──
 # By default npm's global prefix is /usr (writable only by root) so any
-# `pi install npm:<pkg>` or `npm install -g <pkg>` invoked by the
-# developer user would EACCES. Pointing the prefix into ~/.pi places
-# user-installed packages on the named volume, which means they survive
-# container recreation AND image rebuilds.
+# `npm install -g <pkg>` invoked by the developer user would EACCES.
+# Pointing the prefix into ~/.config/opencode places user-installed
+# packages on the devbox-opencode-config named volume, which means they
+# survive container recreation AND image rebuilds.
+#
+# NOTE (v2.0.0): this prefix previously lived at ~/.pi/npm-global — a
+# pi-specific path. With pi removed (see docs/CLEANUP-v2.0.0.md) it now
+# lives under ~/.config/opencode, which is a persistent named volume in
+# BOTH docker-compose.yml and docker-compose.shared.yml (the old ~/.pi
+# volume was only in the former). A one-time migration shim in
+# entrypoint-user.sh copies any existing ~/.pi/npm-global contents to the
+# new prefix on first start so user-installed globals are not lost.
 #
 # IMPORTANT: in this split-build layout the variant Dockerfile inherits
 # this prefix at build time. To keep the baked binaries on /usr (so the
-# ~/.pi volume mount doesn't shadow them), the variant Dockerfile MUST
-# run each `npm install -g` with NPM_CONFIG_PREFIX=/usr in the per-RUN
+# volume mount doesn't shadow them), the variant Dockerfile MUST run each
+# `npm install -g` with NPM_CONFIG_PREFIX=/usr in the per-RUN
 # environment. See Dockerfile.variant.
-ENV NPM_CONFIG_PREFIX=/home/${USER_NAME}/.pi/npm-global
-ENV PATH="/home/${USER_NAME}/.pi/npm-global/bin:${PATH}"
+ENV NPM_CONFIG_PREFIX=/home/${USER_NAME}/.config/opencode/npm-global
+ENV PATH="/home/${USER_NAME}/.config/opencode/npm-global/bin:${PATH}"
 
 # ── Shell defaults (bash history, aliases, readline) ─────────────────
 RUN mkdir -p /etc/skel-devbox
 COPY rootfs/home/developer/.bash_aliases     /etc/skel-devbox/.bash_aliases
 COPY rootfs/home/developer/.inputrc          /etc/skel-devbox/.inputrc
+COPY rootfs/home/developer/.gitignore_global /etc/skel-devbox/.gitignore_global
+
+# ── Editor defaults: system-wide Neovim true-colour ──────────────────
+# /etc/xdg/nvim/sysinit.vim is Neovim's system vimrc: it loads for every user
+# (before any personal ~/.config/nvim) and can still be overridden per-user.
+# Enables termguicolors so the default theme renders in 24-bit colour instead
+# of a muddy 256-colour fallback. Pairs with kitty-terminfo (installed above).
+COPY rootfs/etc/xdg/nvim/sysinit.vim /etc/xdg/nvim/sysinit.vim
+
+# ── Terminal support: xterm-ghostty terminfo alias ──────────────────
+# ncurses-term (installed above) covers wezterm/alacritty/foot/st and the base
+# `ghostty` entry, but Ghostty connects with TERM=xterm-ghostty, for which no
+# distro packages an entry. Ship a thin alias (use=ghostty) and compile it into
+# the system terminfo db with `tic -x`, so it inherits the maintained ghostty
+# capability set. The `infocmp` check fails the build if the entry didn't land.
+COPY rootfs/usr/local/share/terminfo-src/ghostty.terminfo /usr/local/share/terminfo-src/ghostty.terminfo
+RUN tic -x -o /usr/share/terminfo /usr/local/share/terminfo-src/ghostty.terminfo && \
+    infocmp -x xterm-ghostty >/dev/null
 
 # ── Entrypoint ────────────────────────────────────────────────────────
 COPY rootfs/usr/local/lib/opencode-devbox/ /usr/local/lib/opencode-devbox/
+COPY rootfs/usr/local/bin/dot-watch /usr/local/bin/dot-watch
+# Reader for the build manifest baked in Dockerfile.variant. Printed at
+# container start by entrypoint-user.sh; also available on demand.
+COPY rootfs/usr/local/bin/opencode-devbox-version /usr/local/bin/opencode-devbox-version
+# Image-baked skills + harness instruction. Under /usr/local so a named volume
+# over a home dir (e.g. devbox-opencode-config on ~/.config/opencode) can't
+# shadow them; entrypoint-user.sh links them into ~/.agents/skills/ and
+# ~/.config/opencode/instructions/ on every start. See
+# rootfs/usr/local/share/opencode-devbox/skills/VENDORED.md.
+COPY rootfs/usr/local/share/opencode-devbox/ /usr/local/share/opencode-devbox/
 COPY entrypoint.sh /usr/local/bin/entrypoint.sh
 COPY entrypoint-user.sh /usr/local/bin/entrypoint-user.sh
 RUN chmod +x /usr/local/bin/entrypoint.sh /usr/local/bin/entrypoint-user.sh \
+    /usr/local/bin/dot-watch \
+    /usr/local/bin/opencode-devbox-version \
     /usr/local/lib/opencode-devbox/*.py
 
 # Start as root — entrypoint adjusts UID/GID then drops to developer

Dockerfile.variant

@@ -3,32 +3,29 @@
 # FROMs a base-<hash> image produced by Dockerfile.base and adds only
 # the variant-specific tools (opencode, pi, oh-my-opencode-slim, Go).
 #
-# The four published variants are produced from THIS Dockerfile by
+# The two published variants are produced from THIS Dockerfile by
 # varying build args:
 #
-#   variant            INSTALL_OPENCODE  INSTALL_OMOS  INSTALL_PI
-#   ─────────────────  ────────────────  ────────────  ──────────
-#   base               true              false         false
-#   omos               true              true          false
-#   with-pi            true              false         true
-#   omos-with-pi       true              true          true
-#   pi-only            false             false         true
+#   variant   INSTALL_OPENCODE  INSTALL_OMOS
+#   ────────  ────────────────  ────────────
+#   base      true              false
+#   omos      true              true
 #
-# The `pi-only` variant is the single source of truth for the pi-devbox
-# image (pi + companions, no opencode). It exists so pi-devbox can FROM it
-# without inheriting opencode, while the pi install logic stays defined
-# here in one place.
+# pi was removed in v2.0.0 (it had been deprecated since v1.17.2). It now
+# ships from its own self-contained image: joakimp/pi-devbox:latest
+# (https://gitea.jordbo.se/joakimp/pi-devbox). See docs/CLEANUP-v2.0.0.md
+# for the removal history.
 #
 # Pass `--build-arg BASE_IMAGE=<repo>:base-<hash>` to select the base.
 # The CI workflow computes the base hash from Dockerfile.base + rootfs/
 # + entrypoint*.sh and feeds it in.
 #
 # IMPORTANT: the base image sets NPM_CONFIG_PREFIX to
-# /home/developer/.pi/npm-global so runtime `pi install npm:...` and
-# `npm install -g` by the developer user lands on the named volume.
-# At BUILD time we want the baked binaries on /usr so they survive the
-# volume mount. Each `npm install -g` below therefore prefixes the
-# command with `NPM_CONFIG_PREFIX=/usr`.
+# /home/developer/.config/opencode/npm-global so runtime `npm install -g`
+# by the developer user lands on the named volume. At BUILD time we want
+# the baked binaries on /usr so they survive the volume mount. Each
+# `npm install -g` below therefore prefixes the command with
+# `NPM_CONFIG_PREFIX=/usr`.
 
 ARG BASE_IMAGE
 FROM ${BASE_IMAGE}
@@ -42,84 +39,12 @@ ARG USER_NAME=developer
 # edit, so the cache-hit class of bug that bit pi-devbox v0.74.0..
 # v0.75.5 cannot apply here.
 ARG INSTALL_OPENCODE=true
-ARG OPENCODE_VERSION=1.15.13
+ARG OPENCODE_VERSION=1.18.13
 RUN if [ "${INSTALL_OPENCODE}" = "true" ]; then \
       NPM_CONFIG_PREFIX=/usr npm install -g opencode-ai@${OPENCODE_VERSION} && \
       opencode --version ; \
     fi
 
-# ── Optional: pi coding-agent ────────────────────────────────────────
-# pi-toolkit and pi-extensions are cloned into /opt/. entrypoint-user.sh
-# runs each repo's install.sh on container start so symlinks land under
-# ~/.pi/agent/ on the named volume.
-# PI_VERSION should be passed explicitly by CI as a concrete version
-# (resolved from `npm view @earendil-works/pi-coding-agent version`,
-# see .gitea/workflows/docker-publish-split.yml § resolve-versions).
-# The default `latest` is for local dev convenience only — it has a
-# known cache-hit footgun when used in registry-cached CI builds: the
-# resulting build-arg string is byte-identical across builds, the
-# layer-hash is identical, and the registry buildcache silently reuses
-# the layer from whatever pi version was current when the cache was
-# first populated. Currently masked here because OPENCODE_VERSION (a
-# parent layer) bumps every release; will manifest the moment a
-# vN.N.Nb opencode-version-unchanged release ships. See pi-devbox
-# v0.75.5b 2026-05-23 for the discovery + canonical fix.
-ARG INSTALL_PI=false
-ARG PI_VERSION=latest
-ARG PI_TOOLKIT_REF=main
-ARG PI_EXTENSIONS_REF=main
-# pi-fork (fork tool) + pi-observational-memory (recall tool) live on GitHub
-# under elpapi42. Refs default to the tracked branch for local dev; CI resolves
-# them to concrete commit SHAs (see resolve-versions in docker-publish-split.yml)
-# so the build-arg string changes when upstream moves — same registry-buildcache
-# cache-hit footgun the PI_VERSION/OMOS_VERSION pins guard against. The clone
-# helper for these uses `git fetch <ref>` (not `--branch`) so it accepts both
-# branch names and raw commit SHAs.
-ARG PI_FORK_REPO=https://github.com/elpapi42/pi-fork.git
-ARG PI_FORK_REF=master
-ARG PI_OBSMEM_REPO=https://github.com/elpapi42/pi-observational-memory.git
-ARG PI_OBSMEM_REF=master
-RUN if [ "${INSTALL_PI}" = "true" ]; then \
-      set -e && \
-      git_clone_retry() { \
-        url="$1"; ref="$2"; dest="$3"; \
-        for i in 1 2 3 4 5; do \
-          if git clone --depth 1 --branch "$ref" "$url" "$dest"; then return 0; fi; \
-          rm -rf "$dest"; \
-          echo "git clone $url failed (attempt $i/5), retrying in $((i*5))s..."; \
-          sleep $((i*5)); \
-        done; \
-        return 1; \
-      } && \
-      git_fetch_ref() { \
-        url="$1"; ref="$2"; dest="$3"; \
-        rm -rf "$dest"; mkdir -p "$dest"; \
-        git -C "$dest" init -q && git -C "$dest" remote add origin "$url" && \
-        for i in 1 2 3 4 5; do \
-          if git -C "$dest" fetch --depth 1 origin "$ref" && git -C "$dest" checkout -q FETCH_HEAD; then return 0; fi; \
-          echo "git fetch $url@$ref failed (attempt $i/5), retrying in $((i*5))s..."; \
-          sleep $((i*5)); \
-        done; \
-        return 1; \
-      } && \
-      if [ "${PI_VERSION}" = "latest" ]; then \
-        NPM_CONFIG_PREFIX=/usr npm install -g @earendil-works/pi-coding-agent ; \
-      else \
-        NPM_CONFIG_PREFIX=/usr npm install -g @earendil-works/pi-coding-agent@${PI_VERSION} ; \
-      fi && \
-      pi --version && \
-      git_clone_retry https://gitea.jordbo.se/joakimp/pi-toolkit.git "${PI_TOOLKIT_REF}" /opt/pi-toolkit && \
-      git_clone_retry https://gitea.jordbo.se/joakimp/pi-extensions.git "${PI_EXTENSIONS_REF}" /opt/pi-extensions && \
-      git_fetch_ref "${PI_FORK_REPO}"   "${PI_FORK_REF}"   /opt/pi-fork && \
-      git_fetch_ref "${PI_OBSMEM_REPO}" "${PI_OBSMEM_REF}" /opt/pi-observational-memory && \
-      (cd /opt/pi-fork && npm install --omit=dev --no-audit --no-fund) && \
-      (cd /opt/pi-observational-memory && npm install --omit=dev --no-audit --no-fund) && \
-      echo "pi-toolkit at $(cd /opt/pi-toolkit && git rev-parse --short HEAD)" && \
-      echo "pi-extensions at $(cd /opt/pi-extensions && git rev-parse --short HEAD)" && \
-      echo "pi-fork at $(cd /opt/pi-fork && git rev-parse --short HEAD)" && \
-      echo "pi-observational-memory at $(cd /opt/pi-observational-memory && git rev-parse --short HEAD)" ; \
-    fi
-
 # ── Optional: Go ─────────────────────────────────────────────────────
 ARG INSTALL_GO=false
 ARG GO_VERSION=latest
@@ -139,10 +64,13 @@ RUN if [ "${INSTALL_GO}" = "true" ]; then \
 
 # ── Optional: oh-my-opencode-slim (multi-agent orchestration) ────────
 # Installs Bun runtime and the oh-my-opencode-slim npm package.
-# OMOS_VERSION shares the same cache-hit footgun as PI_VERSION when
-# left at the `latest` default in registry-cached CI builds. CI
-# resolves it via `npm view oh-my-opencode-slim version` and passes
-# the concrete value as a build-arg. See PI_VERSION block above.
+# OMOS_VERSION has a cache-hit footgun when left at the `latest` default
+# in registry-cached CI builds: the resulting build-arg string is byte-
+# identical across builds, so the layer-hash is identical, so the
+# registry buildcache silently reuses the layer from whatever omos
+# version was current when the cache was first populated. CI resolves it
+# via `npm view oh-my-opencode-slim version` and passes the concrete
+# value as a build-arg (see resolve-versions in docker-publish-split.yml).
 ARG INSTALL_OMOS=false
 ARG OMOS_VERSION=latest
 RUN if [ "${INSTALL_OMOS}" = "true" ]; then \
@@ -163,4 +91,58 @@ RUN if [ "${INSTALL_OMOS}" = "true" ]; then \
       NPM_CONFIG_PREFIX=/usr npm install -g oh-my-opencode-slim@${OMOS_VERSION}; \
     fi
 
+# ── Build provenance: OCI labels + on-disk manifest ──────────────────
+# These ARGs are declared LAST, immediately before the layer that uses
+# them, so a changing BUILD_DATE / RELEASE_TAG / SOURCE_REVISION never
+# invalidates the expensive npm-install layers above. OPENCODE_VERSION,
+# OMOS_VERSION and INSTALL_OMOS are already in scope from earlier in this
+# stage and need no re-declaration; MEMPALACE_TOOLKIT_REF is consumed in
+# Dockerfile.base, so it is re-declared here only to land in the labels.
+ARG RELEASE_TAG=dev
+ARG BUILD_DATE=
+ARG SOURCE_REVISION=
+ARG MEMPALACE_TOOLKIT_REF=main
+
+LABEL org.opencontainers.image.version="${RELEASE_TAG}" \
+      org.opencontainers.image.revision="${SOURCE_REVISION}" \
+      org.opencontainers.image.created="${BUILD_DATE}" \
+      se.jordbo.opencode-devbox.opencode-version="${OPENCODE_VERSION}" \
+      se.jordbo.opencode-devbox.install-omos="${INSTALL_OMOS}" \
+      se.jordbo.opencode-devbox.omos-version="${OMOS_VERSION}" \
+      se.jordbo.opencode-devbox.mempalace-toolkit-ref="${MEMPALACE_TOOLKIT_REF}"
+
+# The manifest is written from GROUND TRUTH — the live `opencode --version`,
+# the omos package's installed version (when present), and the actual
+# checked-out HEAD of /opt/mempalace-toolkit (cloned in the base) — not
+# merely the intended build-args. That way it also exposes a dependency
+# that silently resolved to something other than the requested value.
+# oh-my-opencode-slim is present only in the omos variant (JSON null
+# otherwise). NOTE: omos is installed under prefix /usr at build time, so
+# we resolve its dir via `npm root -g` with that prefix rather than the
+# runtime NPM_CONFIG_PREFIX the base sets for the developer volume.
+RUN set -e; \
+    mkdir -p /etc/opencode-devbox; \
+    rev() { git -C "$1" rev-parse HEAD 2>/dev/null || echo "unknown"; }; \
+    OPENCODE_V="$(opencode --version 2>/dev/null | head -n1 | tr -d '\r\n')"; \
+    OMOS_REV='null'; \
+    if [ "${INSTALL_OMOS}" = "true" ]; then \
+      OMOS_DIR="$(NPM_CONFIG_PREFIX=/usr npm root -g 2>/dev/null)/oh-my-opencode-slim"; \
+      OMOS_V="$(node -e "process.stdout.write(require('${OMOS_DIR}/package.json').version)" 2>/dev/null || echo unknown)"; \
+      OMOS_REV="\"${OMOS_V}\""; \
+    fi; \
+    { \
+      echo '{'; \
+      echo "  \"release_tag\": \"${RELEASE_TAG}\","; \
+      echo "  \"build_date\": \"${BUILD_DATE}\","; \
+      echo "  \"source_revision\": \"${SOURCE_REVISION}\","; \
+      echo "  \"opencode_version\": \"${OPENCODE_V}\","; \
+      echo "  \"components\": {"; \
+      echo "    \"opencode\": \"${OPENCODE_V}\","; \
+      echo "    \"oh-my-opencode-slim\": ${OMOS_REV},"; \
+      echo "    \"mempalace-toolkit\": \"$(rev /opt/mempalace-toolkit)\""; \
+      echo "  }"; \
+      echo '}'; \
+    } > /etc/opencode-devbox/build-manifest.json; \
+    echo "── build manifest ──"; cat /etc/opencode-devbox/build-manifest.json
+
 # WORKDIR / ENTRYPOINT / CMD inherited from base.

IDEAS.md

@@ -0,0 +1,60 @@
+# Ideas & backlog
+
+A living list of potential improvements for opencode-devbox that are **not yet
+scheduled**. This is intentionally lightweight — a place to park ideas so they
+aren't lost between sessions. When an item ships, describe it in
+[`CHANGELOG.md`](CHANGELOG.md) and remove it from here.
+
+Rough effort tags: 🟢 small · 🟡 medium · 🔴 large. Status: `idea` (unvetted) ·
+`planned` (agreed, not started).
+
+---
+
+## Supply-chain hardening
+
+- 🟡 `planned` — **Pin CI actions to commit SHAs.** The workflows use floating
+  major tags (`actions/checkout@v4`, `docker/build-push-action@v7`,
+  `docker/setup-buildx-action@v4`, `docker/login-action@v3`,
+  `docker/setup-qemu-action@v3`). This is inconsistent with the project's own
+  philosophy of SHA-pinning *content* refs to defeat floating refs. Pin each
+  action to a SHA with a trailing `# vX.Y.Z` comment. Pairs naturally with the
+  renovate item below to keep the pins fresh.
+
+- 🟡 `planned` — **Vulnerability scanning in CI.** No CVE scan runs on the
+  published images today. Add a `trivy image` (or grype) job to
+  `docker-publish-split.yml` after `smoke`. Start non-blocking (report only),
+  then tighten to fail on `HIGH`/`CRITICAL` with an available fix.
+
+- 🟢🟡 `planned` — **Standardize build provenance → buildx SBOM + attestations.**
+  The image already carries hand-rolled provenance (OCI labels +
+  `build-manifest`). `docker/build-push-action` can emit a standard SBOM and
+  SLSA provenance attestation nearly for free (`provenance: mode=max`,
+  `sbom: true`). Makes provenance machine-consumable and pairs well with the
+  trivy item (scan the SBOM).
+
+## Dockerfile hardening
+
+- 🟡 `idea` — **Address hadolint DL4006 properly.** Currently ignored in
+  `.hadolint.yaml`. The clean fix is `SHELL ["/bin/bash", "-o", "pipefail",
+  "-c"]` so piped `RUN`s fail on the first non-zero stage. This changes the
+  default `RUN` shell from `sh` to `bash` for all subsequent layers, so it is
+  base-affecting and needs a careful pass over existing `RUN`s before removing
+  the ignore.
+
+## Developer experience
+
+- 🟢 `idea` — **`Makefile`/`justfile` for local iteration.** Reproducing a CI
+  build locally means hand-assembling many `--build-arg`s. Thin targets
+  (`make build-base`, `make build-variant`, `make smoke`, `make lint`) would
+  make local testing painless and document the canonical invocations.
+
+- 🟡 `idea` — **Dependency-update automation (renovate).** With CI actions
+  SHA-pinned (above), a `renovate.json` keeps those pins — plus the pinned tool
+  versions (`ACTIONLINT_VERSION`, `HADOLINT_VERSION`, gosu, etc.) — current via
+  automated PRs. Requires a renovate runner against the Gitea instance.
+
+## Housekeeping
+
+- 🟢 `idea` — **Registry retention for `base-<hash>` tags.** The base-hash
+  caching scheme accumulates `base-<hash>` tags over time. Confirm whether the
+  registry prunes old ones, and add a retention/cleanup step if not.

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Joakim Persson
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

README.md

@@ -2,6 +2,8 @@
 
 Portable AI developer environment in a Docker container. Run [opencode](https://opencode.ai) on any Docker-capable machine with configurable LLM providers, dev tools, and host filesystem access.
 
+> **Looking for pi?** As of **v2.0.0** the [pi](https://github.com/earendil-works/pi) coding-agent is no longer bundled here. It now ships as its own self-contained image, **[`joakimp/pi-devbox`](https://gitea.jordbo.se/joakimp/pi-devbox)** (also on [Docker Hub](https://hub.docker.com/r/joakimp/pi-devbox)), built on a shared Debian base and able to share this image's mempalace palace. Pull `joakimp/pi-devbox:latest` instead of the old `*-with-pi` / `pi-only` tags. See the [v2.0.0 CHANGELOG](CHANGELOG.md) for the migration details (including the `~/.pi/npm-global` → `~/.config/opencode/npm-global` prefix move).
+
 ## Why?
 
 The official `ghcr.io/anomalyco/opencode` image (now archived) was Alpine-based and minimal — no git, no dev tools, broken PTY support due to musl/glibc incompatibility. This project provides a **Debian-based, production-ready** alternative using the current v1.x release.
@@ -25,7 +27,7 @@ $EDITOR .env
 docker compose run --rm devbox
 ```
 
-This pulls `joakimp/opencode-devbox:latest` from Docker Hub, mounts `WORKSPACE_PATH` at `/workspace`, and drops you straight into opencode. Use `bash` instead of (no command) to land in a shell first — useful for `aws sso login`, `pi`, `omos`, etc.
+This pulls `joakimp/opencode-devbox:latest` from Docker Hub and mounts `WORKSPACE_PATH` at `/workspace`. With no command (as above) the image's default `CMD` (`bash -l`) drops you into a **login shell** — from there run `opencode` to start the harness, or do `aws sso login` first, launch `omos`, etc. To start opencode directly and skip the shell, pass it as the command: `docker compose run --rm devbox opencode`.
 
 **Want to hack on the image itself, follow upstream changes, or rebuild from source?** Clone the repo:
 
@@ -57,6 +59,8 @@ docker compose run --rm devbox
 - **Rust via rustup** — `rustup-init` included; bootstrap Rust on demand with `rustup-init -y`
 - **Optional runtimes** — Python (apt), Go via build args (Node.js always included — required for opencode v1.x)
 - **Multi-agent orchestration** — optional [oh-my-opencode-slim](https://github.com/alvinunreal/oh-my-opencode-slim) integration via build arg
+- **Browser automation** — `agent-browser` + a headless Chromium baked in, so the agent can drive a real browser to *verify* front-end work (live DOM, layout, WebGL) instead of guessing
+- **YAML/JSON tooling** — `jq` plus mikefarah **`yq` v4** (note: replaced Debian's Python `yq` in v2.9.0 — v4 syntax, not jq syntax)
 - **AWS CLI v2** — built-in SSO/Bedrock authentication with headless device-code flow
 - **Multi-arch** — amd64 and arm64
 
@@ -123,7 +127,7 @@ docker compose exec -u developer devbox aws --version
 | Variable | Description | Default |
 |---|---|---|
 | `OPENCODE_PROVIDER` | LLM provider (`anthropic`, `openai`, `amazon-bedrock`) | `anthropic` |
-| `OPENCODE_MODEL` | Model override | Provider default |
+| `OPENCODE_MODEL` | Model override — any `<provider>/<model>` string, written verbatim to the config. See [Choosing a provider and model](#choosing-a-provider-and-model) | Provider default (see below) |
 | `ANTHROPIC_API_KEY` | Anthropic API key | — |
 | `OPENAI_API_KEY` | OpenAI API key | — |
 | `AWS_REGION` | AWS region for Bedrock | `us-east-1` |
@@ -135,18 +139,51 @@ docker compose exec -u developer devbox aws --version
 | `DEVBOX_LAN_ACCESS` | LAN-access mode: `auto` (jump only on VM-backed hosts), `jump` (always), `off` | `auto` |
 | `HOST_SSH_USER` | Username to SSH into the host as (required for the LAN jump) | — |
 | `DEVBOX_HOST_ALIAS` | Hostname used to reach the container host | `host.docker.internal` |
+| `DEVBOX_LAN_AUTOJUMP_PRIVATE` | `1` = ProxyJump *any* RFC1918 (private) IP through the host, so bare `dssh user@<ip>` works on whatever LAN the host is currently on | `0` |
 | `USER_UID` | Override container user UID | Auto-detect from `/workspace` |
 | `USER_GID` | Override container user GID | Auto-detect from `/workspace` |
 | `LANG` | System locale | `en_US.UTF-8` |
 | `LANGUAGE` | Language priority list | `en_US:en` |
 | `LC_ALL` | Override all locale settings | `en_US.UTF-8` |
 | `EDITOR` | Default text editor | `nvim` |
+| `COLORTERM` | Advertises 24-bit colour to colour-aware tools. Export empty (`COLORTERM=`) on a terminal without true-colour support | `truecolor` |
+| `AGENT_BROWSER_EXECUTABLE_PATH` | Chromium binary used by `agent-browser` (a stable symlink into Playwright's versioned browser dir) | `/usr/local/bin/agent-chrome` |
 | `ENABLE_OMOS` | Enable oh-my-opencode-slim multi-agent orchestration | `false` |
 | `OMOS_TMUX` | Enable tmux pane integration for OMOS | `false` |
-| `OMOS_SKILLS` | Install OMOS recommended skills on first run | `true` |
-| `OMOS_RESET` | Force regenerate OMOS config on next start | `false` |
+| `OMOS_SKILLS` | Symlink bundled OMOS skills from the image into `~/.agents/skills/` each start | `true` |
+| `OMOS_RESET` | Force regenerate OMOS config on next start (does not affect skills) | `false` |
+| `OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS` | Enable opencode's native background subagents. Baked on in the image because OMOS V2+ default orchestration depends on it. Set `false` to opt out. opencode marks this **experimental** — see [AGENTS.md](AGENTS.md) removal trigger | `true` |
 | `SKILLSET_CONTAINER_PATH` | Path to skillset repo inside container (for auto-deploy when not at /workspace/skillset) | Auto-detect |
 
+### Choosing a provider and model
+
+The image ships a sensible default model **per provider**, but nothing is hard-wired — you can change it without rebuilding.
+
+| `OPENCODE_PROVIDER` | Default model baked in |
+|---|---|
+| `anthropic` (default) | `anthropic/claude-opus-5` |
+| `amazon-bedrock` | `amazon-bedrock/global.anthropic.claude-opus-5` |
+| `openai` | `openai/gpt-5.6` |
+
+**These are only defaults.** They apply when `OPENCODE_MODEL` is unset, and only for the provider you selected. To use anything else, set `OPENCODE_MODEL` in your `.env`:
+
+```bash
+# A cheaper/faster Anthropic tier
+OPENCODE_MODEL=anthropic/claude-sonnet-5
+
+# A different provider entirely — no baked default needed, the value is
+# written verbatim as the `model` field, so any provider opencode supports works
+OPENCODE_PROVIDER=openai
+OPENCODE_MODEL=openai/gpt-5.6-luna
+```
+
+A few things worth knowing:
+
+- **You are not locked to Anthropic or Bedrock.** The defaults above lean Anthropic only because that's the most common setup here. Set `OPENCODE_PROVIDER` (plus `OPENCODE_MODEL` if the per-provider default isn't what you want) and the Anthropic defaults never come into play.
+- **Defaults only seed a *new* config.** `generate-config.py` never overwrites an existing `~/.config/opencode/opencode.jsonc`, and that file lives on the persistent `devbox-opencode-config` volume — so if you hand-edit the model there, your edit survives restarts *and* image upgrades. Changing `OPENCODE_MODEL` afterwards will **not** rewrite it; edit the config directly, or delete it and let the entrypoint regenerate.
+- **Switching model doesn't require a rebuild** — it's an env var, so `docker compose up -d --force-recreate` (with a fresh config, per the point above) is enough.
+- Defaults are defined in one place: `DEFAULT_MODELS` in `rootfs/usr/local/lib/opencode-devbox/generate-config.py`.
+
 ### Reaching your LAN from the container
 
 The devbox works the same way whether the host is **native Linux Docker** or a **VM-backed** runtime (macOS OrbStack / Docker Desktop, or Docker Desktop on Windows) — but their networking differs:
@@ -154,32 +191,59 @@ The devbox works the same way whether the host is **native Linux Docker** or a *
 - **Native Linux Docker:** the host NATs container egress onto its LAN, so other devices on your LAN are reachable directly. Nothing to configure.
 - **VM-backed (macOS / Docker Desktop):** the container runs in a Linux VM behind the host's network stack. The host's *directly-attached* LAN peers are **not** bridged into the container by default — only the host itself and *routed* subnets are reachable.
 
-On every start the entrypoint detects which case applies. On VM-backed hosts it generates a writable `~/.ssh-local/config` that uses the **host as an SSH jump** to reach LAN peers; on native Linux it does nothing.
+On every start the entrypoint runs `setup-lan-access.sh`, which always writes a writable `~/.ssh-local/config` sidecar. The sidecar does two things regardless of host OS: redirect `ControlPath` into the writable `~/.ssh-local/cm/` (so ControlMaster sockets don't hit the read-only `~/.ssh` bind-mount) and `Include ~/.ssh/config`. On VM-backed hosts it additionally inserts the **SSH-jump-via-host block** so you can reach LAN peers; on native Linux that block is omitted (LAN is reachable directly) but the sidecar is still written — so `dssh`/`dscp` and ControlMaster work on native Linux with a read-only `~/.ssh` too. The jump keypair lives in `~/.ssh-local`, persisted by the `devbox-ssh-local` named volume — generated **once** and reused across container updates.
 
-**To enable it on a VM-backed host:**
+**To enable it on a VM-backed host (one-time setup per machine):**
 
 1. Set `HOST_SSH_USER=<your host username>` in `.env`.
-2. Start the container once. The entrypoint prints a public key — append it to your host's `~/.ssh/authorized_keys`.
+2. Start the container once. When it generates the jump key it prints a ready-to-paste line — run it **on the host** to authorize the key:
+   ```bash
+   echo 'ssh-ed25519 AAAA…devbox-jump@…' >> ~/.ssh/authorized_keys
+   ```
 3. Ensure the host's SSH server is on (on macOS: System Settings → General → Sharing → Remote Login).
-4. Reach the host with `dssh host`, and reach LAN peers by adding `ProxyJump host` to their entries in your bind-mounted `~/.ssh/config`:
+4. Reach the host itself with `dssh host`. (`dssh`/`dscp` wrap `ssh -F ~/.ssh-local/config`.)
+
+Because the key is persisted, you do this **once per machine** — not after every `docker compose up --force-recreate`. You'll only see the authorize line again if you reset the `devbox-ssh-local` volume.
+
+That alone gets you `container → host`. To reach **named LAN peers** by name, give them a `ProxyJump host` override. Don't add it to the shared `~/.ssh/config` entries — the host itself reaches those peers *directly*, and a jump-through-`host` would break the host's own access (and that file is mounted read-only anyway). Instead, drop the overrides in a **host-owned** file that the container Includes ahead of your `~/.ssh/config`:
 
 ```sshconfig
-# in your host ~/.ssh/config (mounted read-only into the container)
-Host my-nas
-    HostName 192.168.1.50
-    User admin
+# ~/.config/devbox-shell/ssh-lan.conf  — on the host, bind-mounted in
+# Only ProxyJump goes here; HostName/User/IdentityFile are inherited
+# (first-value-wins) from the matching block in your ~/.ssh/config.
+Host my-nas pve pbs
     ProxyJump host
 ```
 
-Then `dssh my-nas` routes container → host → LAN peer. (`dssh`/`dscp` wrap `ssh -F ~/.ssh-local/config`; the host config is pulled in via `Include`.)
+Now `dssh my-nas` routes container → host → LAN peer, pulling HostName/User/key from your existing `~/.ssh/config`. See [`ssh-lan.conf.example`](ssh-lan.conf.example).
 
-> This ships the **mechanism** only — your specific target hosts live in your own `~/.ssh/config`, never baked into the image. Set `DEVBOX_LAN_ACCESS=off` to disable, or `=jump` to force it (e.g. native Linux with `extra_hosts: ["host.docker.internal:host-gateway"]`).
+**Roaming / unnamed peers.** Because the jump always targets `host` (= the host on whatever LAN it's currently joined to), you can reach the *current* LAN from anywhere. To make bare `dssh user@<private-ip>` jump automatically without naming peers, set `DEVBOX_LAN_AUTOJUMP_PRIVATE=1` — it ProxyJumps any RFC1918 address through the host. It matches the address you *type* (not the resolved HostName), so named hosts that already carry their own ProxyJump are unaffected.
+
+**Public IPs go direct.** The container has normal internet egress, so a host with a public IP (or one reached via a *public* jump host) connects straight out — the local `host` jump is not involved. e.g. a `Host bastion` whose `HostName` is public, and everything that `ProxyJump bastion`, works from the container by name with no extra setup.
+
+> This ships the **mechanism** only — your specific target hosts are facts about *your* network (and a laptop roams between several), so they live in your own host-side config, never baked into the image. Set `DEVBOX_LAN_ACCESS=off` to disable, or `=jump` to force it (e.g. native Linux with `extra_hosts: ["host.docker.internal:host-gateway"]`).
+
+#### Gotcha: per-host `ControlPath` and read-only `~/.ssh`
+
+The base image bakes a `Host *` default (`/etc/ssh/ssh_config.d/00-devbox-controlmaster.conf`) that points `ControlPath` at the writable, per-container `/tmp/sshcm/` (created mode-700 on every start by `entrypoint-user.sh`). Multiplexing therefore works out of the box. **But your bind-mounted `~/.ssh/config` is read first, and SSH uses the first value it sees** — so any per-host block that sets its own `ControlPath` under `~/.ssh/` (a common CGNAT-multiplexing pattern, e.g. `ControlPath ~/.ssh/cm/%r@%h:%p`) **wins, and then fails inside the container** because `~/.ssh` is mounted **read-only** — the master socket can't bind (`cannot bind … Read-only file system`).
+
+This bites any in-container tool that opens an SSH connection to a remote host (git over SSH, `rsync`, remote-execution agents): the master fails to establish and the connection either errors or silently degrades.
+
+**Fix (host-side, one line):** in your host's `~/.ssh/config`, either drop the per-host `ControlPath` (to inherit the writable baked default) or point it at a path that's writable inside the container too:
+
+```sshconfig
+Host my-remote
+    # was: ControlPath ~/.ssh/cm/%r@%h:%p   ← read-only in the container
+    ControlPath /tmp/sshcm/%r@%h:%p          # writable on both host and container
+```
+
+`/tmp/sshcm/` is also writable on the host (macOS/Linux), so native (non-container) `ssh` from the host keeps working and CGNAT multiplexing is preserved (`ControlMaster`/`ControlPersist` unchanged — only the socket *directory* moves). Note SSH does not create the `ControlPath` parent dir; the container makes `/tmp/sshcm` every start, but on the host run `mkdir -p /tmp/sshcm` once if it doesn't already exist.
 
 ### Custom opencode config
 
-Opencode configuration is persisted automatically via the named volume `devbox-opencode-config`. This volume is mounted at `/home/developer/.config/opencode` by default — no host directory setup required. All changes to `opencode.jsonc`, skills, and (on the OMOS variant) `oh-my-opencode-slim.json` survive container recreation.
+Opencode configuration is persisted automatically via the named volume `devbox-opencode-config`. This volume is mounted at `/home/developer/.config/opencode` by default — no host directory setup required. Changes to `opencode.jsonc` and (on the OMOS variant) `oh-my-opencode-slim.json` survive container recreation. Auto-deployed skills are *not* stored here — skillset and OMOS skills are symlinked into `~/.agents/skills/` and rebuilt on every start (see [Custom skills](#custom-skills) and [docs/omos-skills.md](docs/omos-skills.md)).
 
-When an existing `opencode.jsonc` is found in the volume, the `OPENCODE_PROVIDER` auto-config is skipped.
+When an existing `opencode.jsonc` is found in the volume, the `OPENCODE_PROVIDER` auto-config is skipped — the live config is never overwritten. However, on each start `generate-config.py` checks whether the config it *would* generate for your current environment differs from the live one, and if so writes a **`opencode.jsonc.proposed`** sidecar in the same directory. This is a manual-merge reference only — opencode never loads `.proposed` files. It is removed automatically once your live config matches the current image defaults. A one-line hint is logged when one is present. Differences may reflect new image defaults (e.g. a newly-added MCP server) *or* your own past edits — the file header explains both.
 
 **Alternative: host bind-mount** — if you specifically want to share config from the host (e.g. to version-control it or sync across machines), replace the named volume with a bind mount:
 
@@ -202,6 +266,28 @@ When a skillset repo is detected, its skills are symlinked into `~/.agents/skill
 
 > **Warning:** Do not bind-mount a host `~/.agents/skills` directory directly into the container. This conflicts with the symlink-based auto-deploy mechanism and causes broken skill references.
 
+On the OMOS variant, the five skills bundled with oh-my-opencode-slim are also symlinked into `~/.agents/skills/` on each start — **from the image**, so pulling a newer image updates them with no installer run and no config reset. See [docs/omos-skills.md](docs/omos-skills.md).
+
+#### Image-baked fallback skills (work with no skillset mounted)
+
+Even with **no skillset repo mounted**, the image ships two skills and one
+harness instruction so a fresh container still knows how to behave here:
+
+- `opencode-devbox-environment` and `mempalace` skills are baked under
+  `/usr/local/share/opencode-devbox/skills/` and symlinked into
+  `~/.agents/skills/` on start — **only when a skill of that name is not already
+  present**, so a mounted skillset or an OMOS-bundled skill always wins.
+- `instructions/opencode-devbox.md` is symlinked into
+  `~/.config/opencode/instructions/` (which opencode auto-loads at session
+  start). It points the agent at the `opencode-devbox-environment` skill and the
+  MemPalace continuity protocol. It uses a distinct filename, so it never
+  collides with a mounted skillset's own instructions.
+
+Because the link targets live in the image (not the persistent
+`devbox-opencode-config` volume), `docker compose pull` + recreate refreshes
+them for free. See
+[`rootfs/usr/local/share/opencode-devbox/skills/VENDORED.md`](rootfs/usr/local/share/opencode-devbox/skills/VENDORED.md).
+
 ### Neovim configuration
 
 The image includes neovim 0.12 with `EDITOR=nvim` set by default. To use your own neovim config (and have plugins auto-install via lazy.nvim on first start), mount it from the host:
@@ -211,6 +297,39 @@ volumes:
   - ~/.config/nvim:/home/developer/.config/nvim:ro
 ```
 
+24-bit colour is on by default — a system-wide `sysinit.vim` (in `/etc/xdg/nvim/`) enables `termguicolors` so strings and comments render in full colour over SSH instead of a muddy 256-colour palette. It loads before any personal config and stays overridable per-user (`:set notermguicolors`, or your own `~/.config/nvim/init`).
+
+> **Not a vi person?** The image also ships two non-modal editors alongside nvim:
+> **nano** (classic, minimal) and **micro** (modern — desktop-style `Ctrl+S`/`Ctrl+Q`
+> keys, mouse, syntax highlighting). `EDITOR` stays `nvim`; opt in per-shell with
+> `export EDITOR=nano` (or `micro`), or for git with `git config --global core.editor micro`.
+
+### Document conversion (Markdown → PDF)
+
+The image ships [`pandoc`](https://pandoc.org/) with [`typst`](https://typst.app/) as the PDF engine, so Markdown → PDF works out of the box with no LaTeX install:
+
+```bash
+pandoc README.md -o readme.pdf --pdf-engine=typst
+```
+
+The bundled pandoc typst template defaults the font to `Libertinus Serif`, so a bare `--pdf-engine=typst` renders without needing `-V mainfont`. For higher-fidelity or complex layouts, install TeX Live on demand and use `--pdf-engine=xelatex` instead.
+
+### Browser automation (agent-browser)
+
+The base bakes the [`agent-browser`](https://www.npmjs.com/package/agent-browser) CLI plus a Playwright-managed headless Chromium, so an agent can drive a **real browser** — open pages, click/fill, `eval` JavaScript, snapshot the DOM, take screenshots — and thereby *verify* front-end work rather than assuming it renders correctly. Useful for checking live DOM state, layout, popup positioning, and WebGL.
+
+`AGENT_BROWSER_EXECUTABLE_PATH` is preset to the baked browser, so it works with no setup:
+
+```bash
+agent-browser open https://example.com
+agent-browser screenshot --path /workspace/shot.png
+agent-browser skills get core --full   # full command set, version-matched to the CLI
+```
+
+The browser is resolved through the stable symlink `/usr/local/bin/agent-chrome`, which points into Playwright's per-version, per-architecture browser directory — so image upgrades don't break the path. Playwright's redundant `chromium_headless_shell` build is removed at build time; `agent-browser` drives the full Chromium (headless included).
+
+> **Size note:** Chromium is the single largest thing in the base (~625 MB). It ships in *both* variants because verification is broadly useful. If you maintain a fork and don't need it, drop the `agent-browser` layer from `Dockerfile.base` and lower the smoke-test size thresholds accordingly.
+
 ### Python development with uv
 
 The image includes Python 3.13 (from Debian Trixie) and [uv](https://docs.astral.sh/uv/), a fast Python package manager that replaces pip, venv, and pyenv:
@@ -382,7 +501,7 @@ Enable optional language runtimes, pin a specific opencode version, or lock any
 
 ```bash
 docker compose build --build-arg INSTALL_GO=true
-docker compose build --build-arg OPENCODE_VERSION=1.5.0
+docker compose build --build-arg OPENCODE_VERSION=1.18.13
 docker compose build --build-arg NVIM_VERSION=0.12.1   # pin to a specific version
 ```
 
@@ -390,15 +509,14 @@ docker compose build --build-arg NVIM_VERSION=0.12.1   # pin to a specific versi
 |---|---|---|
 | `INSTALL_GO` | `false` | Go toolchain (resolves latest stable from go.dev when `GO_VERSION=latest`) |
 | `INSTALL_MEMPALACE` | `true` | [MemPalace](https://github.com/MemPalace/mempalace) local AI memory system (~300 MB — disable to shrink image if you don't need MCP memory) |
-| `INSTALL_MEMPALACE_TOOLKIT` | `true` | [mempalace-toolkit](https://gitea.jordbo.se/joakimp/mempalace-toolkit) bash wrappers (`mempalace-session`, `mempalace-docs`). Cloned at build time from `MEMPALACE_TOOLKIT_REF` (default `main`). Requires `INSTALL_MEMPALACE=true`. |
+| `INSTALL_MEMPALACE_TOOLKIT` | `true` | [mempalace-toolkit](https://gitea.jordbo.se/joakimp/mempalace-toolkit) bash wrappers (`mempalace-session`, `mempalace-docs`). Cloned at build time from `MEMPALACE_TOOLKIT_REPO` at ref `MEMPALACE_TOOLKIT_REF` (default `main`). Requires `INSTALL_MEMPALACE=true`. |
 | `INSTALL_OMOS` | `false` | [oh-my-opencode-slim](https://github.com/alvinunreal/oh-my-opencode-slim) multi-agent orchestration (installs Bun and plugin) |
-| `INSTALL_OPENCODE` | `true` | Install opencode. Set `false` to build a pi-only image (still includes Bun if `INSTALL_OMOS=true`; for a fully stripped pi-only image see the `pi-devbox` repo). |
-| `INSTALL_PI` | `false` | Install [pi](https://github.com/earendil-works/pi) as alternative/complementary harness. Both clones [pi-toolkit](https://gitea.jordbo.se/joakimp/pi-toolkit) (~5 MB) and [pi-extensions](https://gitea.jordbo.se/joakimp/pi-extensions) (~1 MB) into `/opt/`; entrypoint deploys them on container start. ~150 MB total image growth. |
-| `PI_VERSION` | `latest` | npm version of `@earendil-works/pi-coding-agent`. Floats by default (image rebuild = pi update). |
-| `PI_TOOLKIT_REF`, `PI_EXTENSIONS_REF` | `main` | Git refs for the toolkit/extensions clones. Pin to a tag/commit for reproducibility. |
+| `INSTALL_OPENCODE` | `true` | Install opencode. Set `false` to build a base with no harness (still includes Bun if `INSTALL_OMOS=true`). |
 | `OPENCODE_VERSION` | *(pinned per release)* | opencode npm version. Drives the image tag and is intentionally not floated. |
 | `NODE_VERSION` | `22` | Node.js major version. Pinned to protect against upstream breaking changes across majors. |
-| `GOSU_VERSION`, `FZF_VERSION`, `GIT_LFS_VERSION`, `NVIM_VERSION`, `BAT_VERSION`, `EZA_VERSION`, `ZOXIDE_VERSION`, `UV_VERSION`, `GITEA_MCP_VERSION`, `GO_VERSION`, `OMOS_VERSION` | `latest` | All GitHub/Gitea/go.dev-hosted binaries resolve to the newest upstream release at build time. Override with a specific version to pin. Resolved versions are logged in CI output. |
+| `MEMPALACE_VERSION` | *(pinned per release)* | MemPalace PyPI version. Deliberately pinned so every bump is a reviewable diff — a past unpinned install swept in an MCP schema regression. Bumped in lockstep with the sibling `pi-devbox` repo. Unrelated to `MEMPALACE_TOOLKIT_REF` despite the shared prefix. |
+| `DEBIAN_VERSION` | `trixie-slim` | OS base image tag. Pinned to a codename; apt resolves updates within that release. |
+| `GOSU_VERSION`, `FZF_VERSION`, `GIT_LFS_VERSION`, `GITLEAKS_VERSION`, `NVIM_VERSION`, `BAT_VERSION`, `EZA_VERSION`, `ZOXIDE_VERSION`, `UV_VERSION`, `GITEA_MCP_VERSION`, `GO_VERSION`, `OMOS_VERSION`, `MICRO_VERSION`, `TEALDEER_VERSION`, `TYPST_VERSION`, `YQ_VERSION`, `AGENT_BROWSER_VERSION`, `PLAYWRIGHT_VERSION` | `latest` | All GitHub/Gitea/go.dev/npm-hosted binaries resolve to the newest upstream release at build time. Override with a specific version to pin. Resolved versions are logged in CI output. |
 
 > **Reproducibility note:** With `latest` defaults, two builds of the same `v{opencode}` tag may embed different tool versions if upstream releases have happened in between. This is intentional — it means every rebuild picks up upstream CVE fixes automatically. If you need a bit-for-bit reproducible build, pass explicit `*_VERSION` args. The CI smoke test logs the resolved versions for every release build.
 
@@ -430,7 +548,7 @@ ENABLE_OMOS=true
 docker compose run --rm devbox
 ```
 
-On first start, the entrypoint runs the oh-my-opencode-slim installer in non-interactive mode. It generates agent configuration at `~/.config/opencode/oh-my-opencode-slim.json` inside the container. The default preset uses OpenAI models — edit the generated config or mount your own to customize.
+On first start, the entrypoint runs the oh-my-opencode-slim installer in non-interactive mode. It generates agent configuration at `~/.config/opencode/oh-my-opencode-slim.json` inside the container. The default preset uses OpenAI models — edit the generated config or mount your own to customize. The installer no longer manages skills (`--skills=no`); the bundled skills are symlinked from the image on every start — see [docs/omos-skills.md](docs/omos-skills.md).
 
 ### OMOS Environment Variables
 
@@ -438,8 +556,8 @@ On first start, the entrypoint runs the oh-my-opencode-slim installer in non-int
 |---|---|---|
 | `ENABLE_OMOS` | `false` | Activate oh-my-opencode-slim on container start |
 | `OMOS_TMUX` | `false` | Enable tmux pane integration (tmux is included in the base image) |
-| `OMOS_SKILLS` | `true` | Install recommended skills (simplify, agent-browser, cartography) |
-| `OMOS_RESET` | `false` | Force regenerate config on next start (backs up existing config) |
+| `OMOS_SKILLS` | `true` | Symlink the bundled OMOS skills (`clonedeps`, `codemap`, `deepwork`, `oh-my-opencode-slim`, `simplify`) from the image into `~/.agents/skills/` on each start. Independent of `ENABLE_OMOS`. See [docs/omos-skills.md](docs/omos-skills.md) |
+| `OMOS_RESET` | `false` | Force regenerate config on next start (backs up existing config). Does **not** affect skills |
 
 ### Custom Configuration
 
@@ -457,79 +575,6 @@ ping all agents
 
 All six agents should respond if your provider authentication is working.
 
-## pi (alternative/complementary harness)
-
-[pi](https://github.com/earendil-works/pi) is a lightweight TUI coding-agent that can run alongside opencode in the same container. Both harnesses share the mempalace install and palace data — wing/diary entries created by one are visible to the other.
-
-### Setup
-
-Pre-built pi-enabled images are available on Docker Hub as `joakimp/opencode-devbox:latest-with-pi` (base + pi) and `joakimp/opencode-devbox:latest-omos-with-pi` (OMOS + pi). Pulling one of those tags is the fastest path. If you want pi **without** opencode, use the separate, leaner [`joakimp/pi-devbox`](https://gitea.jordbo.se/joakimp/pi-devbox) image instead (it's built from the same `Dockerfile.variant` with `INSTALL_OPENCODE=false`, published in its own repo so an opencode-devbox tag never ships without opencode). Alternatively, build from source:
-
-### Build
-
-```bash
-docker compose build --build-arg INSTALL_PI=true
-# Or: pin a pi version
-docker compose build --build-arg INSTALL_PI=true --build-arg PI_VERSION=0.73.0
-# Or: pi-only image (no opencode, smaller)
-docker compose build --build-arg INSTALL_PI=true --build-arg INSTALL_OPENCODE=false
-```
-
-### Run
-
-The default `compose run --rm devbox` invocation drops to a login bash so you can choose:
-
-```bash
-docker compose run --rm devbox       # bash, then `pi` or `opencode` or `aws sso login`
-docker compose run --rm devbox pi     # launch pi directly
-docker compose run --rm devbox opencode
-```
-
-For an attached `compose up -d` container, both harnesses are reachable via `compose exec`:
-
-```bash
-docker compose exec -u developer devbox pi
-docker compose exec -u developer devbox opencode
-docker compose exec -u developer devbox bash
-```
-
-### What gets installed
-
-- **`pi` CLI** — npm-installed globally at build time. Version pinned by `PI_VERSION`.
-- **pi-toolkit** — keybindings.json (mosh/tmux newline fixes), pi-env.zsh (AWS env loader), settings.json template. Cloned to `/opt/pi-toolkit`; deployed to `~/.pi/agent/` on first container start.
-- **pi-extensions** — 7 extensions, cloned to `/opt/pi-extensions` and symlinked into `~/.pi/agent/extensions/`:
-  - `confirm-destructive` — confirm-prompt before dangerous bash commands and session actions.
-  - `ext-toggle` — `/ext` slash command to list and enable/disable extensions at runtime (rename-to-disable; survives `/reload`).
-  - `git-checkpoint` — per-turn `git stash` checkpoint, restorable on `/fork`.
-  - `mcp-loader` — generic MCP server loader. Reads an `mcp` block from `~/.pi/agent/settings.json` (same shape as opencode and Claude Desktop) and connects to each declared server, exposing the tools as native pi tools. Supports both **local stdio** subprocesses (`uvx mcp-searxng`, `gitea-mcp`, …) and **remote streamable-HTTP** servers per MCP spec 2025-03-26 (e.g. `https://mcp.context7.com/mcp`). Adds a `/mcp` slash command for runtime status / toggle (same UX as `/ext`). See [`pi-extensions/AGENTS.md`](https://gitea.jordbo.se/joakimp/pi-extensions/src/branch/main/AGENTS.md) for transport details and the `headers` config for auth tokens.
-  - `notify` — native terminal notification when the agent finishes.
-  - `ssh-controlmaster` — transparent SSH remote execution via persistent ControlMaster socket (when pi is launched with `--ssh user@host`).
-  - `todo` — `todo` tool for the agent + `/todos` for the user.
-- **mempalace bridge** — separate `mempalace.ts` extension symlinked from the cloned `mempalace-toolkit`. Provides pi's MCP tools for palace search/diary/knowledge-graph with bespoke agent-identity injection from `$MEMPALACE_AGENT_NAME`. Coexists with `mcp-loader` rather than replacing it — don't list `mempalace` in settings.json's `mcp` block too, or you'll get duplicate tool registrations.
-- **MCP servers (none baked in beyond mempalace)** — the loader registers nothing by default. Add servers by editing `~/.pi/agent/settings.json` and `/reload`. Examples (mcp-searxng for web search, context7 for live library docs) are in the `pi-extensions` README.
-
-### Persistence
-
-`~/.pi/` is mounted on the `devbox-pi-config` named volume. Everything below survives container recreate **and** image rebuilds:
-
-- `~/.pi/agent/settings.json` (provider/model, theme selection, the `mcp` block, and the `packages` array tracking installed pi packages).
-- `~/.pi/agent/extensions/` (hand-placed extensions and the symlinks deployed by `pi-extensions/install.sh`).
-- `~/.pi/agent/sessions/`, `~/.pi/agent/auth.json`.
-- `~/.pi/agent/git/<host>/<path>/` (pi packages installed via `pi install git:...`).
-- `~/.pi/npm-global/` (pi packages installed via `pi install npm:...`, plus any `npm install -g` invoked as the `developer` user). `NPM_CONFIG_PREFIX` is pre-set in the image, the prefix's `bin/` is on `PATH`, and the directory itself lives on the volume — so user-installed themes, skills, and extensions survive everything short of `docker compose down -v`.
-
-The **baked** pi binary (and pi-toolkit / pi-extensions repos under `/opt/`) live on the image filesystem, not the volume. Image rebuild is the upgrade path for those — same contract as `OPENCODE_VERSION`. If you `npm install -g @earendil-works/pi-coding-agent` yourself, the user-installed copy on the volume wins via `PATH` order and survives image rebuilds.
-
-### Configuration
-
-The entrypoint copies `pi-toolkit/settings.example.json` to `~/.pi/agent/settings.json` on first start. Edit it to set provider/model:
-
-```bash
-docker compose exec -u developer devbox $EDITOR ~/.pi/agent/settings.json
-```
-
-The AWS env loader (`pi-env.zsh`) reads `~/.config/pi/.env` if you bind-mount one; otherwise pi uses container env vars passed via `.env`.
-
 ## AWS Bedrock Authentication
 
 When using AWS Bedrock as your LLM provider, you need:
@@ -613,6 +658,28 @@ Add mempalace as an MCP server in your `opencode.jsonc` (inside `~/.config/openc
 
 This gives opencode access to 29 MCP tools for searching memory, querying the knowledge graph, managing wings/rooms/drawers, and agent diaries.
 
+### Sharing one palace across containers (external server)
+
+By default each container runs its own local `mempalace-mcp`. To share **one**
+palace across several containers/harnesses (pi + opencode + native), run the
+bundled server and point clients at it:
+
+```bash
+docker compose -f docker-compose.mempalace.yml up -d      # serves mempalace-mcp over HTTP
+```
+
+Then set in each client's `.env`:
+
+```bash
+MEMPALACE_REMOTE_URL=http://<reachable-host>:8765/mcp
+# MEMPALACE_REMOTE_TOKEN=...   # optional; sent as Authorization: Bearer
+```
+
+`generate-config.py` then registers mempalace as a `remote` MCP endpoint instead
+of the local stdio command — no local `mempalace-mcp` is spawned and the
+`devbox-palace` volume is irrelevant. ⚠ The HTTP transport is **unauthenticated**;
+keep it on a trusted network or behind a reverse proxy that enforces the token.
+
 ### Basic usage
 
 ```bash
@@ -647,7 +714,66 @@ Both wrappers are idempotent and dedup-aware — re-running them on unchanged in
 
 For weekly automated runs, the toolkit ships ready-to-use scheduler templates (systemd user timer, launchd user agent, cron) in its [`contrib/`](https://gitea.jordbo.se/joakimp/mempalace-toolkit/src/branch/main/contrib) directory. The `*-devbox` variants are designed for this container: host-side schedulers that `docker exec` into the running opencode-devbox.
 
-Disable the toolkit (keeps mempalace itself) with `--build-arg INSTALL_MEMPALACE_TOOLKIT=false`. Pin to a specific ref with `--build-arg MEMPALACE_TOOLKIT_REF=v0.3.0` once tagged releases exist.
+Disable the toolkit (keeps mempalace itself) with `--build-arg INSTALL_MEMPALACE_TOOLKIT=false`. Pin to a specific ref with `--build-arg MEMPALACE_TOOLKIT_REF=v0.3.0` once tagged releases exist. Repoint the clone URL with `--build-arg MEMPALACE_TOOLKIT_REPO=<url>` for forked or air-gapped builds (see below).
+
+### Building a fork / relocated build
+
+The canonical build clones `mempalace-toolkit` from `gitea.jordbo.se`. That URL is an overridable build-arg (defaulting to the canonical origin), so a fork or a build on a host that can't reach that gitea can repoint it at a mirror, another host, or a local `file://` path **without editing the Dockerfiles**:
+
+| Build-arg | Default | Dockerfile |
+|---|---|---|
+| `MEMPALACE_TOOLKIT_REPO` | `https://gitea.jordbo.se/joakimp/mempalace-toolkit.git` | base |
+
+Each companion also has a matching `*_REF` arg (branch name or commit SHA). Example — build against a forked mempalace-toolkit:
+
+```bash
+# base first
+docker build -f Dockerfile.base -t myorg/opencode-devbox:base-dev \
+  --build-arg MEMPALACE_TOOLKIT_REPO=https://github.com/myorg/mempalace-toolkit.git .
+
+# then the variant FROM that base
+docker build -f Dockerfile.variant -t myorg/opencode-devbox:dev \
+  --build-arg BASE_IMAGE=myorg/opencode-devbox:base-dev \
+  --build-arg OPENCODE_VERSION=1.17.20 .
+```
+
+Note: mempalace-toolkit clones anonymously (no token needed). Only the `resolve-versions` CI job calls the gitea API (which needs a token for public repos). A plain `docker build` like the above skips that job entirely, so no credentials are required.
+
+Provenance build-args (all optional; populate the OCI labels and `/etc/opencode-devbox/build-manifest.json` — see below): `RELEASE_TAG`, `BUILD_DATE`, `SOURCE_REVISION`. CI sets these automatically; a manual build leaves them at harmless defaults.
+
+### Build provenance (labels + manifest)
+
+Every published image is self-describing. Inspect the OCI labels without pulling the filesystem:
+
+```bash
+docker inspect --format '{{json .Config.Labels}}' joakimp/opencode-devbox:latest | jq .
+```
+
+`org.opencontainers.image.{version,revision,created}` plus `se.jordbo.opencode-devbox.{opencode-version,install-omos,omos-version,mempalace-toolkit-ref}` record the intended opencode version, omos status, and companion refs. The on-disk `/etc/opencode-devbox/build-manifest.json` records **ground truth** — the live `opencode --version`, the installed `oh-my-opencode-slim` version (or `null` in the base variant), and the actual checked-out HEAD of `/opt/mempalace-toolkit` — so a tag is reconstructable after CI logs rotate:
+
+```bash
+docker run --rm --entrypoint= joakimp/opencode-devbox:latest cat /etc/opencode-devbox/build-manifest.json
+```
+
+From **inside** a running container, `opencode-devbox-version` reads that manifest for you — and it is printed automatically as the first line of output when the container starts, so "which image am I in?" is answered before you ask:
+
+```bash
+opencode-devbox-version           # human-readable summary
+opencode-devbox-version --json    # raw manifest, for scripting
+opencode-devbox-version --quiet   # one line: "v2.9.0 (a1b2c3d)"
+```
+
+```text
+opencode-devbox v2.9.0
+  built:    2026-08-04T12:00:00Z  (source a1b2c3d4e5f6)
+  opencode: 1.18.13
+  components:
+    opencode: 1.18.13
+    oh-my-opencode-slim: 2.2.9
+    mempalace-toolkit: 0123456789ab
+```
+
+It also performs a **drift check**: because `npm install -g` as the `developer` user lands on the persistent config volume, a locally-installed `opencode` can shadow the baked one. If the live version differs from the baked one, the command says so instead of silently reporting the manifest's value.
 
 ### Storage
 
@@ -718,6 +844,7 @@ Defaults you get out of the box:
 - **Aliases** — `ls`/`ll`/`la` use `eza`, `cat` uses `bat`, `gs`/`gd`/`gl` for git, safe `rm`/`mv`/`cp`.
 - **Integrations** — `zoxide` (`z <fragment>` to jump), `fzf` Ctrl-R / Ctrl-T key bindings.
 - **Prompt marker** — `[devbox]` prefix so it's always obvious you're inside the container.
+- **Host SSH reachability check** — on the first shell of a freshly (re)created container, a one-time probe checks whether the Docker host is reachable over SSH. If it isn't (host Remote Login off, or the container's `devbox_jump` key not yet authorized), it prints a clear two-step fix with the container's public key inline; it stays silent once SSH works and is skipped on later shells.
 
 ### Overriding the defaults
 
@@ -751,6 +878,16 @@ cp /etc/skel-devbox/.bash_aliases ~/.bash_aliases
 rm ~/.bash_aliases
 ```
 
+## Terminal compatibility
+
+The base ships `ncurses-term` and `kitty-terminfo` on top of the default `ncurses-base`, plus a compiled `xterm-ghostty` alias, so modern terminal emulators resolve their `TERM` correctly over SSH instead of degrading to a dumb fallback. Covered out of the box: WezTerm, Alacritty, foot, st, kitty (`xterm-kitty`), Ghostty (`xterm-ghostty`), and iTerm2 / xterm (`xterm-256color`).
+
+`COLORTERM=truecolor` is also baked in, so colour-aware tools (Neovim's auto-detect, `bat`, `delta`) render in 24-bit colour instead of falling back to 256 colours. Pairs with the system-wide Neovim `termguicolors` default. If you connect from a terminal that lacks true-colour support, export `COLORTERM=` (empty) to opt out.
+
+## Global gitignore
+
+The image bakes a `~/.gitignore_global` and wires it via `git config --global core.excludesFile`, so personal/tooling artifacts are ignored across every repo in the container without per-repo `.gitignore` entries. Seeded patterns include `*.bak`, `*.bak.*`, `*~`, `*.orig`, `*.swp`, `*.tmp`, and `**/.claude/settings.local.json` (Claude Code's per-machine settings, which can carry credentials). It is seeded only if absent — edit it freely, and your version survives recreate — and the `core.excludesFile` wiring is skipped if you already set one.
+
 ## Secret Scanning
 
 A [gitleaks](https://github.com/gitleaks/gitleaks) pre-commit hook prevents accidentally committing API keys, passwords, or other secrets.
@@ -792,8 +929,12 @@ Container (Debian trixie)
 ├── opencode binary
 ├── oh-my-opencode-slim (optional — multi-agent orchestration plugin, includes Bun)
 ├── AWS CLI v2 (SSO + Bedrock auth)
-├── neovim 0.12, tmux, htop, bat, eza, zoxide, uv, rustup, make, gcc, g++, rsync
-├── git, git-crypt, age, gitleaks, ssh, ripgrep, fd, fzf, jq, curl, tree
+├── neovim 0.12, nano, micro, tmux, htop, bat, eza, zoxide, uv, rustup, make, gcc, g++, rsync
+├── git, git-crypt, age, gitleaks, ssh, ripgrep, fd, fzf, jq, yq, curl, tree
+├── pandoc + typst (Markdown→PDF via --pdf-engine=typst; HTML/etc.)
+├── graphviz (dot diagram rendering)
+├── tldr (tealdeer — quick command examples)
+├── dot-watch (auto-render .dot to PNG on save)
 ├── Node.js (for MCP servers)
 ├── Bun (optional — included with oh-my-opencode-slim)
 ├── entrypoint.sh (UID adjustment, git config, provider setup)
@@ -822,4 +963,4 @@ Container (Debian trixie)
 
 ## License
 
-MIT
+MIT — see [`LICENSE`](LICENSE). The MIT license covers this repository's own contents (Dockerfiles, entrypoint scripts, `rootfs/` seeds, CI workflows, docs). The **published container images** additionally *bundle* third-party software, each under its own terms — see [`THIRD_PARTY.md`](THIRD_PARTY.md) for the summary.

THIRD_PARTY.md

@@ -0,0 +1,51 @@
+# Third-party notices
+
+opencode-devbox is distributed under the MIT License (see [`LICENSE`](LICENSE)),
+which covers **this repository's own contents** — the Dockerfiles, entrypoint
+scripts, `rootfs/` seeds, CI workflows, and docs.
+
+The **published container images** (`joakimp/opencode-devbox:*`) additionally
+*bundle* third-party software, each of which remains under its own license. This
+file is a good-faith summary; the authoritative sources are the upstream
+projects and, for OS packages, the per-package copyright files inside the image
+at `/usr/share/doc/<package>/copyright`.
+
+## opencode and agent components (installed in the variant layer)
+
+| Component | Upstream | License |
+| --- | --- | --- |
+| opencode | github.com/sst/opencode | MIT |
+| oh-my-opencode-slim (OMOS) *(`-omos` variant only)* | npm | MIT |
+| mempalace / mempalace-toolkit | authored by the maintainer (Joakim Persson) | MIT |
+| gitea-mcp | gitea.com/gitea/gitea-mcp | MIT |
+
+## Tooling baked into the base image
+
+| Component | Upstream | License (best effort) |
+| --- | --- | --- |
+| gosu | github.com/tianon/gosu | Apache-2.0 |
+| Node.js | nodejs.org | MIT (bundles components under their own licenses) |
+| uv | github.com/astral-sh/uv | Apache-2.0 OR MIT |
+| Neovim | neovim.io | Apache-2.0 + Vim license |
+| micro | github.com/zyedidia/micro | MIT |
+| Pandoc | pandoc.org | GPL-2.0-or-later |
+| Typst | github.com/typst/typst | Apache-2.0 |
+| Graphviz | graphviz.org | CPL-1.0 |
+| ripgrep / fd / bat / eza / zoxide / tealdeer / yq (mikefarah) | respective repos | MIT / Apache-2.0 / Unlicense (varies) |
+| agent-browser | npmjs.com/package/agent-browser | see package |
+| Playwright | github.com/microsoft/playwright | Apache-2.0 |
+| Chromium *(fetched by Playwright)* | chromium.org | BSD-3-Clause + others |
+| bun *(`-omos` variant only)* | github.com/oven-sh/bun | MIT |
+
+## Base OS
+
+The image is built `FROM` a Debian base and installs packages via `apt`. Debian
+and its packages are distributed under their respective licenses (GPL, LGPL,
+MIT, BSD, and others). See each package's copyright file in the image under
+`/usr/share/doc/<package>/copyright`.
+
+---
+
+*Licenses marked "best effort" are widely known but were not each verified at
+the exact bundled version; consult the upstream project for authoritative
+terms. Corrections welcome.*

docker-compose.mempalace.yml

@@ -0,0 +1,84 @@
+# Shared MemPalace server (optional) — one palace for many clients.
+#
+# Runs `mempalace-mcp` over HTTP so several containers/harnesses (pi +
+# opencode + native) can share ONE palace instead of each keeping its own.
+# Point every client at it by setting, in that client's .env:
+#
+#   MEMPALACE_REMOTE_URL=http://<reachable-host>:8765/mcp
+#
+# (see .env.example). When set, the client connects over HTTP and does NOT
+# spawn its own local mempalace-mcp.
+#
+# Start:   docker compose -f docker-compose.mempalace.yml up -d
+# Stop:    docker compose -f docker-compose.mempalace.yml down
+# Logs:    docker compose -f docker-compose.mempalace.yml logs -f
+#
+# Why reuse the devbox image? mempalace-mcp is already installed in it, and
+# reusing it GUARANTEES the server's mempalace version matches the clients'
+# (both are pinned by the same image build). Override with a slimmer image via
+# MEMPALACE_SERVER_IMAGE if you prefer (it must provide `mempalace-mcp`).
+#
+# ⚠ SECURITY: mempalace-mcp's HTTP transport has NO authentication of its own.
+# Do NOT expose port 8765 to an untrusted network. The default below binds to
+# 127.0.0.1 (host loopback) only. To let sibling containers reach it, either
+# attach them to the shared `mempalace-net` network (container-to-container, no
+# host port needed — use http://mempalace-server:8765/mcp), or front it with a
+# reverse proxy that enforces MEMPALACE_REMOTE_TOKEN as `Authorization: Bearer`.
+
+name: mempalace-server
+
+services:
+  mempalace:
+    image: ${MEMPALACE_SERVER_IMAGE:-joakimp/opencode-devbox:latest}
+    container_name: mempalace-server
+    # Bypass the devbox entrypoint (dev-shell/LAN/config setup) and run the
+    # HTTP MCP server directly. HOME + explicit --palace pin the data path so
+    # it does not depend on the image's default user/HOME. Runs as root so it
+    # can initialise the fresh named volume; the volume is dedicated to this
+    # server (clients reach it over HTTP, never by mounting it).
+    entrypoint: []
+    user: "0:0"
+    environment:
+      - HOME=/data
+    command:
+      - mempalace-mcp
+      - --transport
+      - http
+      - --host
+      - "0.0.0.0"
+      - --port
+      - "8765"
+      - --palace
+      - /data/.mempalace
+    restart: unless-stopped
+    # Loopback-only by default (see SECURITY note). Use "8765:8765" to expose on
+    # all host interfaces, or drop `ports:` entirely and rely on mempalace-net.
+    ports:
+      - "127.0.0.1:8765:8765"
+    volumes:
+      # The shared palace data — precious; back this up.
+      - mempalace-shared:/data/.mempalace
+      # Embedding-model cache (~79 MB, disposable) so search does not re-download.
+      - mempalace-shared-chroma:/data/.cache/chroma
+    networks:
+      - mempalace-net
+    healthcheck:
+      # A tools/list round-trip proves the server is answering MCP (python3 is
+      # always present — mempalace itself is a python tool in the image).
+      test:
+        - CMD
+        - python3
+        - -c
+        - "import urllib.request,json; d=json.dumps({'jsonrpc':'2.0','id':1,'method':'tools/list','params':{}}).encode(); r=urllib.request.Request('http://127.0.0.1:8765/mcp',data=d,headers={'Content-Type':'application/json','Accept':'application/json'}); urllib.request.urlopen(r,timeout=5).read()"
+      interval: 30s
+      timeout: 10s
+      retries: 3
+      start_period: 60s
+
+volumes:
+  mempalace-shared:
+  mempalace-shared-chroma:
+
+networks:
+  mempalace-net:
+    name: mempalace-net

docker-compose.shared.yml

@@ -38,6 +38,10 @@ services:
       - TERM=xterm-256color
       - GITEA_ACCESS_TOKEN=${GITEA_ACCESS_TOKEN:-}
       - GITEA_HOST=${GITEA_HOST:-}
+      # MemPalace transport: empty = LOCAL per-container palace (default);
+      # set MEMPALACE_REMOTE_URL to share one external palace. See .env.example.
+      - MEMPALACE_REMOTE_URL=${MEMPALACE_REMOTE_URL:-}
+      - MEMPALACE_REMOTE_TOKEN=${MEMPALACE_REMOTE_TOKEN:-}
     volumes:
       # Host workspace — user's project directory
       - ${WORKSPACE_PATH:-~/src}:/workspace
@@ -73,7 +77,9 @@ services:
       # Persist uv data (Python installs)
       - devbox-uv:/home/developer/.local/share/uv
 
-      # Optional: persist MemPalace data (conversation memory, knowledge graph)
+      # Optional: persist MemPalace data (conversation memory, knowledge graph).
+      # LOCAL palace only (default). In EXTERNAL mode (MEMPALACE_REMOTE_URL set)
+      # the shared server owns the data, so this volume is irrelevant.
       # - devbox-palace:/home/developer/.mempalace
 
       # Optional: persist ChromaDB embedding model cache (~79 MB)

docker-compose.yml

@@ -25,8 +25,6 @@ services:
     #   args:
     #     INSTALL_GO: "false"
     #     INSTALL_OMOS: "false"
-    #     INSTALL_PI: "false"
-    #     # PI_VERSION: "latest"
     #     # INSTALL_OPENCODE: "true"
     container_name: opencode-devbox
     stdin_open: true
@@ -38,6 +36,12 @@ services:
       - GITHUB_PERSONAL_ACCESS_TOKEN=${GITHUB_PERSONAL_ACCESS_TOKEN:-}
       - GITEA_ACCESS_TOKEN=${GITEA_ACCESS_TOKEN:-}
       - GITEA_HOST=${GITEA_HOST:-}
+      # MemPalace transport: leave empty for a LOCAL per-container palace
+      # (default). Set MEMPALACE_REMOTE_URL to share ONE external MemPalace
+      # across containers/harnesses; MEMPALACE_REMOTE_TOKEN is an optional
+      # bearer token. See .env.example.
+      - MEMPALACE_REMOTE_URL=${MEMPALACE_REMOTE_URL:-}
+      - MEMPALACE_REMOTE_TOKEN=${MEMPALACE_REMOTE_TOKEN:-}
     volumes:
       # Host workspace — mount your project here
       - ${WORKSPACE_PATH:-.}:/workspace
@@ -58,7 +62,18 @@ services:
       # the container's skill/instruction symlinks independent from the host,
       # allowing both native and containerized opencode on the same machine.
       - devbox-opencode-config:/home/developer/.config/opencode
-      - devbox-pi-config:/home/developer/.pi
+      # Legacy pi config volume (pi was removed in v2.0.0). Left commented so
+      # fresh installs carry no dead weight. If you are UPGRADING from a
+      # pre-v2.0.0 image and had global npm packages installed via
+      # `npm install -g` (they lived under ~/.pi/npm-global), uncomment this
+      # for ONE container start: entrypoint-user.sh's migration shim copies
+      # them to ~/.config/opencode/npm-global, after which you can remove it.
+      # - devbox-pi-config:/home/developer/.pi
+      # Persist the generated LAN-jump keypair (~/.ssh-local) across recreates.
+      # setup-lan-access.sh generates this key once and reuses it; persisting
+      # it means you authorize it on the host ONCE rather than re-authorizing
+      # after every `docker compose up --force-recreate`.
+      - devbox-ssh-local:/home/developer/.ssh-local
 
       # NOTE: Do NOT bind-mount ~/.agents/skills/ from the host. The
       # container manages its own skills directory independently — the
@@ -95,6 +110,14 @@ services:
       # - ~/.bash_aliases:/home/developer/.bash_aliases:ro
        # - ~/.inputrc:/home/developer/.inputrc:ro
 
+      # Optional: host-owned shell config + LAN jump overrides (recommended
+      # over the single-file ~/.bash_aliases mount above — it's a directory,
+      # so it survives editors' atomic-save). The image's ~/.bash_aliases
+      # sources ~/.config/devbox-shell/bash_aliases if present, and
+      # setup-lan-access.sh reads ~/.config/devbox-shell/ssh-lan.conf for
+      # named-peer `ProxyJump host` overrides (see ssh-lan.conf.example).
+      # - ~/.config/devbox-shell:/home/developer/.config/devbox-shell:ro
+
       # Optional: persist uv data (Python installs, tool installs)
       # Without this, 'uv python install' must be re-run after container removal.
       - devbox-uv:/home/developer/.local/share/uv
@@ -111,7 +134,10 @@ services:
       - devbox-nvim-data:/home/developer/.local/share/nvim
 
       # Optional: persist MemPalace data (conversation memory, knowledge graph,
-      # embeddings). Without this, palace data is lost on container recreation.
+      # embeddings). Applies to the LOCAL palace only (the default, when
+      # MEMPALACE_REMOTE_URL is unset). Without this, palace data is lost on
+      # container recreation. In EXTERNAL mode the shared server owns the data,
+      # so this volume is irrelevant.
       # - devbox-palace:/home/developer/.mempalace
 
       # Optional: persist ChromaDB embedding model cache (~79 MB, downloaded on
@@ -125,7 +151,8 @@ services:
 
 volumes:
   devbox-opencode-config:
-  devbox-pi-config:
+  # devbox-pi-config:  # legacy (pi removed v2.0.0) — uncomment with the mount above only to migrate old global npm packages
+  devbox-ssh-local:
   devbox-data:
   devbox-state:
   devbox-shell-history:

docs/CLEANUP-v2.0.0.md

@@ -0,0 +1,247 @@
+# PR-5: Retire pi from opencode-devbox
+
+After pi-devbox has shipped v1.0.0 as a fully independent image (with
+its own base + variant Dockerfiles, CI, smoke tests, and docs), the
+pi-related paths in opencode-devbox become dead weight. This PR
+removes them.
+
+## Pre-conditions before merging
+
+This PR should land **only after** all of the following are stable:
+
+1. `pi-devbox v1.0.0` published, smoke tests passing, in active use
+   for at least one release cycle.
+2. Anyone consuming `joakimp/pi-devbox:base-pi-only` directly (e.g.
+   forks pinned to it) has been notified and migrated.
+3. The deprecation warning (PR-1 of this work — see below) has been
+   live for at least one release cycle so consumers have visible
+   notice.
+
+## Files / sections to remove from opencode-devbox
+
+### `Dockerfile.variant`
+
+Remove these blocks entirely:
+
+- The `INSTALL_PI` / `PI_VERSION` / `PI_TOOLKIT_REF` /
+  `PI_EXTENSIONS_REF` / `PI_FORK_REPO` / `PI_FORK_REF` /
+  `PI_OBSMEM_REPO` / `PI_OBSMEM_REF` build-args.
+- The `RUN if [ "${INSTALL_PI}" = "true" ]; then ...` block (entire
+  block — git_clone_retry, git_fetch_ref, npm install
+  pi-coding-agent, the four /opt/pi-* clones, the npm installs in
+  /opt/pi-fork and /opt/pi-observational-memory, and the four
+  rev-parse echoes).
+- All comments referencing pi-only as "the single source of truth for
+  the pi-devbox image" (the variant matrix table, the explanatory
+  paragraph, and the "rationale" comments at the top of the file
+  about pi-only existing for pi-devbox to FROM).
+
+Update the variant matrix table at the top of `Dockerfile.variant`:
+
+```
+   variant            INSTALL_OPENCODE  INSTALL_OMOS
+   ─────────────────  ────────────────  ────────────
+   base               true              false
+   omos               true              true
+```
+
+(only two variants now; pi-only and the with-pi/omos-with-pi axis are
+gone).
+
+### `entrypoint-user.sh`
+
+Remove:
+
+- The pi-toolkit and pi-extensions install hooks (the section that
+  runs `(cd /opt/pi-toolkit && ./install.sh --yes)` etc.).
+- The `~/.pi/agent/settings.json` seeding from
+  `/opt/pi-toolkit/settings.example.json`.
+- Any other pi-conditional blocks (search for `INSTALL_PI`, `pi-toolkit`,
+  `pi-extensions`, `~/.pi/`).
+
+Verify that the AWS Bedrock auth bootstrap (the pi-toolkit AWS env
+loader) is not relied on by opencode users. If it is, lift it out of
+the pi-toolkit dependency (it's small and self-contained).
+
+### `Dockerfile.base`
+
+Remove:
+
+- The `mkdir -p /home/${USER_NAME}/.pi/agent/extensions` line in the
+  standard-directories block. Replace with the equivalent opencode-
+  specific paths if any aren't already present (`~/.config/opencode`
+  is already there).
+- `NPM_CONFIG_PREFIX=/home/${USER_NAME}/.pi/npm-global` — change to
+  `/home/${USER_NAME}/.config/opencode/npm-global` or a more neutral
+  path. Update the corresponding `PATH` env var.
+
+Also update the long base-image header comment to remove the
+"variants for pi-devbox" rationale.
+
+### CI (`.gitea/workflows/docker-publish-split.yml` or equivalent)
+
+Remove:
+
+- The `pi-only` variant build job.
+- The `with-pi` and `omos-with-pi` variant build jobs (they're
+  redundant with the standalone pi-devbox now).
+- The `base-pi-only` tag publish step (which pushes to
+  `joakimp/pi-devbox:base-pi-only` from this repo).
+- The `resolve-pi-version` job step (no longer needed).
+- Smoke-test invocations with `--variant pi-only`, `--variant with-pi`,
+  `--variant omos-with-pi`.
+
+Remaining variants in CI: `base`, `omos`. The "with-pi" axis is
+fully retired.
+
+### `scripts/smoke-test.sh`
+
+Remove:
+
+- The `--variant pi-only`, `--variant with-pi`, `--variant
+  omos-with-pi` branches.
+- pi-related assertions: `pi --version`, the
+  `~/.pi/agent/extensions/*.ts ≥ 4` check, the mempalace.ts bridge
+  gate (mempalace itself stays, but its bridge into pi is no longer
+  this image's concern).
+
+Remaining variant axis in smoke tests: `base`, `omos`.
+
+### `README.md` (and `AGENTS.md`, `DOCKER_HUB.md`)
+
+- Remove the pi-only variant from the "Image variants" table.
+- Remove the with-pi / omos-with-pi variants if they were documented.
+- Remove all sections about pi-toolkit, pi-extensions, pi-fork,
+  pi-observational-memory, ~/.pi paths, and pi-related env vars.
+- Remove the "this image also produces base-pi-only for pi-devbox"
+  notes.
+- Add a single-paragraph **"Looking for pi?"** section pointing to
+  `joakimp/pi-devbox`.
+
+### `Dockerfile` references in `pi-devbox` repo (cleanup of cross-repo coupling)
+
+This isn't a change to opencode-devbox, but it's part of the same
+deprecation:
+
+- Once pi-devbox v1.0.0 is the single source of truth, remove
+  pi-devbox/Dockerfile (the 5-line shim with the long
+  `joakimp/pi-devbox:base-pi-only` rationale comment). It's replaced
+  by `Dockerfile.base` + `Dockerfile.variant` produced by PR-1 of
+  this work.
+
+### Purge the orphaned `base-pi-only*` Hub tags (manual, post-merge)
+
+Until PR-5, the `build-variant-pi-only` job re-publishes
+`joakimp/pi-devbox:base-pi-only` (floating) and a fresh
+`base-pi-only-vX.Y.Z` on **every** opencode-devbox release. These tags
+are orphaned legacy artifacts: pi-devbox v1.0.0+ builds its own
+debian-based `base-<hash>` and **no pi-devbox build input references
+`base-pi-only`** (verified 2026-06-10 — only historical mentions remain
+in pi-devbox's CHANGELOG/AGENTS/DOCKER_HUB). Nothing consumes them.
+
+Once PR-5 deletes the `build-variant-pi-only` job, the publisher is
+gone — so this is the moment to purge the accumulated tags from the
+**`joakimp/pi-devbox`** Docker Hub repo (NOT opencode-devbox):
+
+- Delete the floating `base-pi-only` tag.
+- Delete every versioned snapshot: `base-pi-only-v1.17.2`,
+  `base-pi-only-v1.16.2`, `base-pi-only-v1.15.13e`, … (all of them).
+
+Do this **after** PR-5 is merged and the first post-PR-5 release has
+built, to confirm no new `base-pi-only*` tag reappears. If you purge
+before PR-5, the next opencode-devbox release simply recreates the
+floating tag (whack-a-mole). Deletion is via the Hub UI
+(`hub.docker.com/r/joakimp/pi-devbox/tags`) or the Hub API
+(`DELETE /v2/repositories/joakimp/pi-devbox/tags/<tag>/` with a Hub PAT).
+
+## Two-step deprecation path (recommended)
+
+Rather than a single big-bang removal, use a deprecation cycle:
+
+### Step 1 — pre-PR (lands at the same time as pi-devbox v1.0.0)
+
+Add a deprecation warning to opencode-devbox:
+
+1. **Build-time message** — when `INSTALL_PI=true`,
+   `INSTALL_PI_DEPRECATED=warn` is the default; the variant build
+   prints to stderr:
+   ```
+   ===========================================================
+   DEPRECATION WARNING: INSTALL_PI is deprecated in opencode-devbox
+   and will be removed in v2.0.0. Use joakimp/pi-devbox:latest
+   instead. See https://gitea.jordbo.se/joakimp/pi-devbox
+   ===========================================================
+   ```
+2. **CHANGELOG** entry on opencode-devbox: "INSTALL_PI build-arg path
+   deprecated; will be removed in v2.0.0."
+3. **README and DOCKER_HUB** updates: mark `pi-only`, `with-pi`,
+   `omos-with-pi` variants as deprecated, point to pi-devbox.
+4. The `base-pi-only` tag continues to be published but with a
+   notice in the description: "Internal artifact for pi-devbox.
+   Deprecated; pull joakimp/pi-devbox:latest directly."
+
+### Step 2 — removal PR (this document)
+
+Lands one release cycle (or one calendar month, whichever is later)
+after step 1. Removes everything listed in the per-file sections
+above. Tagged as opencode-devbox v2.0.0 (the major bump signals the
+breaking change).
+
+## Risk assessment
+
+### What could go wrong
+
+- **Someone is consuming `base-pi-only` directly** without going
+  through pi-devbox. The deprecation warning + one-cycle delay should
+  surface this.
+- **Mempalace bridge in pi-extensions** — this stays in pi-devbox; no
+  impact on opencode-devbox.
+- **Shared base assumptions** — opencode-devbox's
+  `~/.pi/npm-global` NPM_CONFIG_PREFIX was a pi-specific design. In
+  the cleanup we move it to a neutral path. Existing opencode-devbox
+  users get a one-time migration: their `npm install -g` packages
+  installed at the old path stop being on PATH. Document this in the
+  v2.0.0 changelog and add a one-liner that copies the old prefix
+  contents to the new one if the old one exists.
+
+### What's safe
+
+- The base apt set, the Go-binary installs, MemPalace, the SSH
+  ControlMaster setup, the entrypoint UID/GID dance — all of these
+  stay. They're not pi-specific.
+- The `omos` variant — fully unaffected.
+- Existing opencode-only users — no change to their workflow.
+
+## Verification
+
+After PR-5 lands, the following should be true:
+
+- `grep -ri "INSTALL_PI\|pi-toolkit\|pi-extensions\|pi-fork\|pi-observational-memory\|base-pi-only" .` in opencode-devbox returns no matches.
+- `docker history joakimp/opencode-devbox:latest` shows no pi-related layers.
+- The opencode-devbox CI matrix builds only `base` and `omos` variants.
+- A post-PR-5 release does NOT recreate any `joakimp/pi-devbox:base-pi-only*` tag (publisher confirmed gone), after which those orphaned tags are purged from the pi-devbox Hub repo.
+- pi-devbox CI is unaffected (it's a different repo).
+- Both repos build cleanly in their own CI without referencing the other.
+
+## Estimated effort
+
+- Step 1 (deprecation warnings): ~2 hours.
+- Step 2 (removal): ~4 hours including local testing of opencode-only
+  build paths.
+- One release cycle of monitoring between them.
+
+Total: ~1 working day of focused effort, spread over a calendar month.
+
+## Order in the broader plan
+
+1. PR-1 on pi-devbox — copy base + variant Dockerfiles, strip
+   opencode/omos paths, tag v1.0.0.
+2. PR-2 on pi-devbox — add pandoc, graphviz, imagemagick, tldr, yq.
+3. PR-3 on pi-devbox — add `:latest-studio` variant.
+4. (Optional) PR-4 on pi-devbox — add `:latest-studio-tex` variant.
+5. PR-pre on opencode-devbox — deprecation warnings (step 1 above).
+6. **PR-5 on opencode-devbox — actual removal (this document, step 2).**
+
+PRs 1–4 are independent and can land in any order on pi-devbox. PR-pre
+should land alongside or shortly after pi-devbox v1.0.0 (PR-1) so
+consumers know to migrate. PR-5 lands one release cycle after PR-pre.

docs/manual-host-publish.md

@@ -29,14 +29,11 @@ cp docs/manual-host-publish.sh /tmp/manual-publish-vX.Y.Z.sh
 # Edit at top of file:
 #   RELEASE_TAG="vX.Y.Z"
 #   BASE_HASH="<12-char hash from CI's base-decide step>"
-#   PI_VERSION="<from npm registry, see step 2 below>"
 #   OMOS_VERSION="<from npm registry, see step 2 below>"
 bash /tmp/manual-publish-vX.Y.Z.sh
 ```
 
-Keep the historical script in `docs/` as-is — it's an archive of the v1.15.12 publish, useful as a reference if a future debug needs to compare exact arg sets across releases. Don't edit it in place.
-
-The sections below explain what the script does and what you need to know to edit those four constants safely.
+The sections below explain what the script does and what you need to know to edit those three constants safely.
 
 ## 1. Pin RELEASE_TAG
 
@@ -49,16 +46,15 @@ git describe --tags --exact-match HEAD
 
 The script asserts `HEAD == ${RELEASE_TAG}^{commit}` before doing anything destructive. If you've drifted, fix it with `git checkout` before running.
 
-## 2. Pin PI_VERSION and OMOS_VERSION
+## 2. Pin OMOS_VERSION
 
-Gitea CI's `resolve-versions` job queries the npm registry at workflow time and threads concrete versions through every variant build, mitigating the silent same-bytes-across-releases regression class documented in `AGENTS.md`. Do the same by hand:
+Gitea CI's `resolve-versions` job queries the npm registry at workflow time and threads the concrete version through the omos variant build, mitigating the silent same-bytes-across-releases regression class documented in `AGENTS.md`. Do the same by hand:
 
 ```bash
-curl -sf https://registry.npmjs.org/@earendil-works%2Fpi-coding-agent/latest | jq -r .version
 curl -sf https://registry.npmjs.org/oh-my-opencode-slim/latest | jq -r .version
 ```
 
-Paste the two version strings into the script's `PI_VERSION` / `OMOS_VERSION` constants. Don't leave the script defaulting to `latest` — the registry buildcache will silently reuse a stale layer if the build-arg byte-equals a previous build.
+Paste the version string into the script's `OMOS_VERSION` constant. Don't leave the script defaulting to `latest` — the registry buildcache will silently reuse a stale layer if the build-arg byte-equals a previous build.
 
 ## 3. Pin BASE_HASH
 
@@ -101,8 +97,8 @@ After the constants are set, the script runs a 5-step procedure. No editing need
 1. **Preflight** — buildx present, tag exists, `HEAD == tag`, multi-arch builder created if missing.
 2. **Base build (conditional)** — probe `${IMAGE}:base-${BASE_HASH}` on the Hub; if missing, build it multi-arch and push. **No `--cache-from` / `--cache-to`.** That's the whole point of this escape. If the base push itself fails the same way CI did, stop — the regression has spread to image push and you need a different host or account, not this runbook.
 3. **Promote `base-latest`** — `docker buildx imagetools create` re-tags by manifest reference. No rebuild.
-4. **Variants × 5** — sequential (not parallel; one host's egress can't saturate five multi-arch pushes safely). Each variant is `Dockerfile.variant` `FROM ${IMAGE}:base-${BASE_HASH}` plus the appropriate `INSTALL_OPENCODE` / `INSTALL_OMOS` / `INSTALL_PI` build-args, tagged `${RELEASE_TAG}${suffix}` and `latest${suffix}`.
-5. **Verify** — prints the digest of all 12 expected tags (10 variant + base-hash + base-latest). Spot-check that each `vX.Y.Z*` and its `latest*` alias share a digest.
+4. **Variants × 2** — sequential (not parallel; one host's egress can't saturate multiple multi-arch pushes safely). Each variant is `Dockerfile.variant` `FROM ${IMAGE}:base-${BASE_HASH}` plus the appropriate `INSTALL_OPENCODE` / `INSTALL_OMOS` build-args, tagged `${RELEASE_TAG}${suffix}` and `latest${suffix}`.
+5. **Verify** — prints the digest of all 6 expected tags (4 variant + base-hash + base-latest). Spot-check that each `vX.Y.Z*` and its `latest*` alias share a digest.
 
 Expected wall time on a recent Mac: ~25-40 min (base ~3 min if rebuilt, each variant ~3-7 min mostly QEMU arm64 emulation).
 

docs/manual-host-publish.sh

@@ -1,35 +1,34 @@
 #!/usr/bin/env bash
-# Manual publish of opencode-devbox v1.15.12 — bypasses broken Gitea-runner
-# Hub push by building & pushing from a developer host (Orbstack/Docker Desktop).
+# Manual publish of opencode-devbox — bypasses a broken Gitea-runner Hub push
+# by building & pushing from a developer host (Orbstack/Docker Desktop).
 #
 # Mirrors what .gitea/workflows/docker-publish-split.yml would do:
 #   1. Build & push Dockerfile.base    → joakimp/opencode-devbox:base-<hash>
 #   2. Promote                         → joakimp/opencode-devbox:base-latest
-#   3. Build & push 5 variants on top of base-<hash>:
-#        :v1.15.12              :latest              (INSTALL_OPENCODE only)
-#        :v1.15.12-omos         :latest-omos         (+ OMOS)
-#        :v1.15.12-with-pi      :latest-with-pi      (+ pi)
-#        :v1.15.12-omos-with-pi :latest-omos-with-pi (+ both)
-#        :v1.15.12-pi-only      :latest-pi-only      (pi, no opencode)
+#   3. Build & push 2 variants on top of base-<hash>:
+#        :vX.Y.Z              :latest              (INSTALL_OPENCODE only)
+#        :vX.Y.Z-omos         :latest-omos         (+ OMOS)
+#
+# pi was removed in v2.0.0 — there are no pi variants here anymore.
 #
 # Usage on your host:
 #   1. Make sure Orbstack/Docker Desktop is running with multi-arch enabled
 #      (docker buildx ls should show linux/amd64,linux/arm64).
 #   2. docker login docker.io   (joakimp account)
-#   3. cd ~/path/to/opencode-devbox && git fetch && git checkout v1.15.12
-#   4. bash /path/to/this/script.sh
+#   3. cd ~/path/to/opencode-devbox && git fetch && git checkout <RELEASE_TAG>
+#   4. Edit RELEASE_TAG / BASE_HASH / OMOS_VERSION below to match the release.
+#   5. bash /path/to/this/script.sh
 #
-# Total expected time: ~25-40 min on a recent Mac (4 multi-arch builds, base
+# Total expected time: ~15-25 min on a recent Mac (2 multi-arch builds, base
 # layers cache after the first variant).
 
 set -euo pipefail
 
 IMAGE="joakimp/opencode-devbox"
-RELEASE_TAG="v1.15.12"
-BASE_HASH="8d72a9e44796"        # sha256 of Dockerfile.base + rootfs/* + entrypoints (computed by CI logic)
+RELEASE_TAG="v2.0.0"            # EDIT per release
+BASE_HASH="REPLACE_ME"          # sha256 of Dockerfile.base + rootfs/* + entrypoints (computed by CI logic)
 BASE_TAG="base-${BASE_HASH}"
-PI_VERSION="0.76.0"             # resolved from npm @earendil-works/pi-coding-agent latest (2026-05-28)
-OMOS_VERSION="1.1.1"            # resolved from npm oh-my-opencode-slim latest (2026-05-28)
+OMOS_VERSION="latest"           # resolve from npm oh-my-opencode-slim latest, then pin
 PLATFORMS="linux/amd64,linux/arm64"
 
 # -------- preflight --------
@@ -65,14 +64,12 @@ fi
 echo "==> [2/7] Promote ${IMAGE}:${BASE_TAG} → ${IMAGE}:base-latest"
 docker buildx imagetools create -t "${IMAGE}:base-latest" "${IMAGE}:${BASE_TAG}"
 
-# -------- 3-5. variants --------
+# -------- 3-4. variants --------
 build_variant() {
-  local suffix="$1"      # "" | "-omos" | "-with-pi" | "-omos-with-pi" | "-pi-only"
+  local suffix="$1"      # "" | "-omos"
   local install_omos="$2"
-  local install_pi="$3"
-  local install_opencode="${4:-true}"
+  local install_opencode="${3:-true}"
   local extra_args=()
-  [[ "$install_pi"   == "true" ]] && extra_args+=(--build-arg "PI_VERSION=${PI_VERSION}")
   [[ "$install_omos" == "true" ]] && extra_args+=(--build-arg "OMOS_VERSION=${OMOS_VERSION}")
 
   local versioned="${IMAGE}:${RELEASE_TAG}${suffix}"
@@ -85,7 +82,6 @@ build_variant() {
     --build-arg "BASE_IMAGE=${IMAGE}:${BASE_TAG}" \
     --build-arg "INSTALL_OPENCODE=${install_opencode}" \
     --build-arg "INSTALL_OMOS=${install_omos}" \
-    --build-arg "INSTALL_PI=${install_pi}" \
     ${extra_args[@]+"${extra_args[@]}"} \
     -t "${versioned}" \
     -t "${floating}" \
@@ -93,29 +89,17 @@ build_variant() {
     .
 }
 
-echo "==> [3/7] Variant: base (opencode only)"
-build_variant ""               false false
+echo "==> [3/4] Variant: base (opencode only)"
+build_variant ""               false
 
-echo "==> [4/7] Variant: omos"
-build_variant "-omos"          true  false
-
-echo "==> [5/7] Variant: with-pi"
-build_variant "-with-pi"       false true
-
-echo "==> [6/7] Variant: omos-with-pi"
-build_variant "-omos-with-pi"  true  true
-
-echo "==> [7/7] Variant: pi-only (pi without opencode)"
-build_variant "-pi-only"       false true  false
+echo "==> [4/4] Variant: omos"
+build_variant "-omos"          true
 
 echo
 echo "==> Done. Verifying tags on Hub:"
 for t in \
   "${RELEASE_TAG}" "latest" \
   "${RELEASE_TAG}-omos" "latest-omos" \
-  "${RELEASE_TAG}-with-pi" "latest-with-pi" \
-  "${RELEASE_TAG}-omos-with-pi" "latest-omos-with-pi" \
-  "${RELEASE_TAG}-pi-only" "latest-pi-only" \
   "${BASE_TAG}" "base-latest"
 do
   d=$(docker manifest inspect "${IMAGE}:${t}" 2>/dev/null | python3 -c "import json,sys,hashlib; m=json.load(sys.stdin); print(m.get('digest','-'))" 2>/dev/null || echo "MISSING")

docs/omos-skills.md

@@ -0,0 +1,110 @@
+# OMOS bundled skills
+
+How the five skills bundled with [oh-my-opencode-slim](https://github.com/alvinunreal/oh-my-opencode-slim)
+(OMOS) are deployed in this image, why the mechanism changed, and how to keep
+them up to date.
+
+## TL;DR
+
+- OMOS bundles five skills: `clonedeps`, `codemap`, `deepwork`,
+  `oh-my-opencode-slim`, `simplify`.
+- They are **symlinked from the image** into `~/.agents/skills/` on every
+  container start by `entrypoint-user.sh` (gated by `OMOS_SKILLS`, default
+  `true`).
+- **To update them, pull a newer image and recreate the container** — no
+  installer run, no config reset:
+  ```bash
+  docker compose pull
+  docker compose up -d --force-recreate
+  ```
+
+## Why it works this way
+
+The skills ship inside the OMOS npm package, baked into the image at
+`/usr/lib/node_modules/oh-my-opencode-slim/src/skills/<name>/`. On start the
+entrypoint creates one absolute symlink per skill into `~/.agents/skills/` —
+the same flat directory the [skillset](#relationship-to-skillset) deploy uses
+and which opencode scans (directly and via the `~/.claude/skills` pointer).
+
+Because the symlink target lives in the **image**, not in a volume, the skill
+content tracks whatever image you run. Pull a newer `*-omos` image and the
+skills update on the next container start. `~/.agents/skills/` is itself an
+ephemeral container-layer directory rebuilt from scratch on every start, so the
+reconcile is idempotent and self-healing.
+
+The symlinks are **absolute** (unlike skillset's relative links): the target is
+always inside the container at a fixed `/usr` path, so there is no host/container
+path divergence to guard against.
+
+## The old mechanism (and the trap it created)
+
+Previously the OMOS *installer* (`oh-my-opencode-slim install --skills=yes`)
+**copied** the skills into `~/.config/opencode/skills/` — but only on the very
+first container start, gated by the absence of `oh-my-opencode-slim.json`. That
+directory is the persistent `devbox-opencode-config` named volume.
+
+The consequence: once the config existed, the installer was skipped forever, so
+the copied skills **froze** at whatever the image shipped on first run. Pulling
+a newer image did nothing (the copies lived in the volume, not the image). The
+only refresh path was `OMOS_RESET=true`, which runs `install --reset` and
+**also overwrites your hand-tuned `opencode.jsonc`** (model choices, agent
+config). Updating skills meant clobbering unrelated config — a bad trade.
+
+The installer has no skills-only refresh flag (`--skills=yes|no` is all-or-
+nothing within a full install; `--reset` overwrites everything), so the fix was
+to stop using the installer for skills entirely. The two `install` invocations
+in `entrypoint-user.sh` now pass `--skills=no`; the installer manages only the
+agent config (`oh-my-opencode-slim.json`).
+
+## One-time migration of frozen copies
+
+Existing volumes still contain the old frozen real directories under
+`~/.config/opencode/skills/`. Because opencode prefers a name found there over
+the same name in `~/.agents/skills/`, those stale copies would *shadow* the
+fresh image-sourced symlinks. On the first start after upgrading, the entrypoint
+**backs each of them up — never deletes** — to
+`~/.config/opencode/skills/<name>.bak.<epoch>` and writes a marker at
+`~/.config/opencode/.omos-skills-migrated` so the migration runs exactly once.
+Only real directories are touched; any symlink in that directory is left alone.
+
+If everything looks right after a few sessions, the backups are safe to remove:
+
+```bash
+rm -rf ~/.config/opencode/skills/*.bak.*
+```
+
+## Relationship to skillset
+
+The [skillset](https://gitea.jordbo.se/joakimp/skillset) repo deploys its own
+version-controlled skills into `~/.agents/skills/` via
+`deploy-skills.sh --bootstrap --prune-stale`, which the entrypoint runs on every
+start — *before* the OMOS reconcile. If a skill name exists in both, **OMOS
+wins**: the reconcile uses `ln -sfn`, which replaces any existing symlink. Today
+the only overlap is `simplify` (the OMOS copy is richer — it bundles
+`codemap.md` and a `README.md`), so `simplify` was removed from the skillset
+repo and the image copy owns the name.
+
+## Configuration
+
+| Variable | Default | Effect |
+|---|---|---|
+| `OMOS_SKILLS` | `true` | Symlink the bundled skills into `~/.agents/skills/` on each start. Set `false` to deploy no skills from the image. |
+
+`OMOS_SKILLS` is **independent of `ENABLE_OMOS`**: the skills are useful to plain
+opencode even when the multi-agent orchestration config is not enabled. The
+reconcile is additionally gated on the bundled-skills source being present, so
+it is automatically a no-op on the non-OMOS (`base`) image variant.
+
+## Troubleshooting
+
+- **A skill didn't update after `docker compose pull`.** Make sure you
+  recreated the container (`docker compose up -d --force-recreate`); a plain
+  restart reuses the old container layer. Confirm the link target —
+  `readlink ~/.agents/skills/deepwork` should point under
+  `/usr/lib/node_modules/oh-my-opencode-slim/src/skills/`.
+- **A skill disappeared.** If OMOS upstream restructured its package the symlink
+  target may no longer exist. The build-time smoke test asserts the source path,
+  so this should be caught in CI; if you hit it at runtime, look for a dangling
+  link in `ls -l ~/.agents/skills/`.
+- **I want a frozen copy back.** It's at
+  `~/.config/opencode/skills/<name>.bak.<epoch>` until you delete it.

docs/plan-lan-access-and-pi-extensions.md

@@ -219,7 +219,8 @@ Add (for `with-pi`/`omos-with-pi`/pi-devbox):
 2. **Latest-but-pinned**: track latest (master HEAD), resolve to SHA in CI build-arg. ✓
 3. **Refactor**: pi-devbox/Dockerfile -> `FROM` the with-pi variant; pi-install in ONE place. ✓
 4. **LAN default** `DEVBOX_LAN_ACCESS=auto`: generate config + print authorize hint when
-   `HOST_SSH_USER` unset; silent no-op on native Linux. ✓
+   `HOST_SSH_USER` unset; silent no-op on native Linux. ✓ *(v2.2.0: sidecar now rendered on
+   all OSes; native Linux no longer skipped — jump block still omitted there)*
 5. **No `DEVBOX_LAN_HOSTS`**: rely on user's bind-mounted `~/.ssh/config` (`ProxyJump host`). ✓
 
 ## Remaining verify-before-merge items

entrypoint-user.sh

@@ -1,6 +1,14 @@
 #!/usr/bin/env bash
 set -euo pipefail
 
+# ── Startup banner: which opencode-devbox build is this? ───────────
+# Printed FIRST, before the setup noise below, so it's the first thing visible
+# when the container starts (CMD is `bash -l`, tty:true in compose, so this
+# reaches the same stream as the interactive shell the user lands in). Reads the
+# ground-truth manifest baked in Dockerfile.variant; a no-op with a short stderr
+# notice on images built before it existed.
+command -v opencode-devbox-version >/dev/null 2>&1 && opencode-devbox-version || true
+
 # ── SSH ControlMaster socket dir ────────────────────────────────
 # Companion to /etc/ssh/ssh_config.d/00-devbox-controlmaster.conf in the
 # base image — that file declares ControlPath=/tmp/sshcm/%r@%h:%p; this
@@ -12,12 +20,16 @@ set -euo pipefail
 mkdir -p /tmp/sshcm
 chmod 700 /tmp/sshcm
 
-# ── LAN access: generic host-OS-agnostic reachability helper ────────
-# On VM-backed hosts (macOS OrbStack / Docker Desktop) the container can't
-# reach the host's directly-attached LAN peers by default; this generates a
-# writable ~/.ssh-local/config that uses the host as an SSH jump. On native
-# Linux (LAN reachable directly) it is a no-op. Controlled by DEVBOX_LAN_ACCESS
-# (auto|jump|off) + HOST_SSH_USER. Always non-fatal. See the script header.
+# ── LAN access + writable SSH sidecar: host-OS-agnostic helper ──────
+# Generates the writable ~/.ssh-local/config on EVERY host OS: a `Host *`
+# ControlPath redirect into ~/.ssh-local/cm (so `ssh -F` / dssh / dscp work
+# even when ~/.ssh is bind-mounted read-only) plus `Include ~/.ssh/config`. On
+# VM-backed hosts (macOS OrbStack / Docker Desktop) it ALSO adds an
+# SSH-jump-via-host block so the container can reach the host's
+# directly-attached LAN peers; on native Linux (LAN reachable directly) the
+# jump block is omitted but the sidecar is still rendered. Controlled by
+# DEVBOX_LAN_ACCESS (auto|jump|off) + HOST_SSH_USER. Always non-fatal. See the
+# script header.
 if [ -r /usr/local/lib/opencode-devbox/setup-lan-access.sh ]; then
   bash /usr/local/lib/opencode-devbox/setup-lan-access.sh || true
 fi
@@ -29,13 +41,39 @@ fi
 # directly.
 SKEL_DIR="/etc/skel-devbox"
 if [ -d "$SKEL_DIR" ]; then
-  for f in .bash_aliases .inputrc; do
+  for f in .bash_aliases .inputrc .gitignore_global; do
     if [ -f "$SKEL_DIR/$f" ] && [ ! -e "$HOME/$f" ]; then
       cp "$SKEL_DIR/$f" "$HOME/$f"
     fi
   done
 fi
 
+# ── v2.0.0 migration: relocate npm global prefix off the legacy ~/.pi path ──
+# Pre-v2.0.0 images set NPM_CONFIG_PREFIX=~/.pi/npm-global (a pi-specific
+# path). v2.0.0 removed pi and moved the prefix to
+# ~/.config/opencode/npm-global, which is a persistent named volume in every
+# compose layout (the old ~/.pi volume was only in docker-compose.yml). If a
+# user upgraded with the old ~/.pi volume still mounted, copy their
+# previously globally-installed npm packages to the new prefix once so they
+# remain on PATH. The marker keeps this idempotent and a no-op for fresh
+# installs; the whole block is harmless when the old path is absent.
+NEW_NPM_PREFIX="$HOME/.config/opencode/npm-global"
+OLD_NPM_PREFIX="$HOME/.pi/npm-global"
+MIGRATION_MARKER="$NEW_NPM_PREFIX/.migrated-from-dot-pi"
+if [ -d "$OLD_NPM_PREFIX" ] && [ ! -f "$MIGRATION_MARKER" ]; then
+  echo "Migrating npm global prefix: ~/.pi/npm-global -> ~/.config/opencode/npm-global"
+  mkdir -p "$NEW_NPM_PREFIX"
+  # cp -n: never overwrite a file already in the new prefix (a freshly
+  # installed package wins over the legacy copy).
+  for sub in lib bin share; do
+    if [ -d "$OLD_NPM_PREFIX/$sub" ]; then
+      mkdir -p "$NEW_NPM_PREFIX/$sub"
+      cp -an "$OLD_NPM_PREFIX/$sub/." "$NEW_NPM_PREFIX/$sub/" 2>/dev/null || true
+    fi
+  done
+  touch "$MIGRATION_MARKER" 2>/dev/null || true
+fi
+
 # ── MemPalace: initialize palace for the workspace if mempalace is installed
 # Creates the palace directory structure on first run. Idempotent — skips
 # if palace already exists, so upgrades from older versions preserve
@@ -62,70 +100,22 @@ fi
 if [ -n "${GIT_USER_EMAIL:-}" ] && ! git config --global user.email &>/dev/null; then
   git config --global user.email "$GIT_USER_EMAIL"
 fi
+# Global gitignore for personal/tooling artifacts (*.bak, *~, *.orig, ...).
+# Seeded above into $HOME/.gitignore_global from /etc/skel-devbox. Point git at
+# it only if the user has not already set their own core.excludesFile.
+if [ -f "$HOME/.gitignore_global" ] && ! git config --global core.excludesFile &>/dev/null; then
+  git config --global core.excludesFile "$HOME/.gitignore_global"
+fi
 
 # ── Generate opencode config from env vars if no config mounted ──────
 # Delegated to a standalone Python script for clarity and testability.
-# The script is idempotent: it never overwrites an existing opencode.json
-# (bind-mounted from host, persisted in named volume, or previously
-# generated) and no-ops if OPENCODE_PROVIDER is unset.
+# The script never overwrites an existing opencode.json/.jsonc (bind-mounted
+# from host, persisted in named volume, or previously generated) and no-ops if
+# OPENCODE_PROVIDER is unset. When a config already exists it instead writes a
+# NON-loaded opencode.jsonc.proposed sidecar (only when newer image defaults
+# differ) for manual review/merge.
 python3 /usr/local/lib/opencode-devbox/generate-config.py
 
-# ── pi: deploy toolkit + extensions + mempalace bridge ─────────────
-# Runs only when pi was baked into the image (INSTALL_PI=true at build).
-# Each install.sh is idempotent and backs up real files before linking,
-# so re-running across container restarts is safe.
-#
-# Order: pi-toolkit first (creates ~/.pi/agent/keybindings.json symlink
-# and writes the AWS env loader), then pi-extensions (symlinks our 6
-# extensions), then settings.json bootstrap from the toolkit template,
-# then the mempalace bridge symlink (one-liner; mempalace-toolkit's
-# install_skill is intentionally skipped to avoid racing with skillset
-# auto-deploy below).
-if command -v pi &>/dev/null; then
-  if [ -d /opt/pi-toolkit ]; then
-    (cd /opt/pi-toolkit && ./install.sh --yes) || \
-      echo "WARN: pi-toolkit install.sh failed (continuing)"
-  fi
-
-  if [ -d /opt/pi-extensions ]; then
-    (cd /opt/pi-extensions && ./install.sh --yes) || \
-      echo "WARN: pi-extensions install.sh failed (continuing)"
-  fi
-
-  # Bootstrap settings.json from template if absent (pi rewrites this
-  # file at runtime — lastChangelogVersion, etc — so we can't symlink it).
-  if [ ! -f "$HOME/.pi/agent/settings.json" ] && \
-     [ -f /opt/pi-toolkit/settings.example.json ]; then
-    cp /opt/pi-toolkit/settings.example.json "$HOME/.pi/agent/settings.json"
-  fi
-
-  # pi↔mempalace MCP bridge — single extension symlink.
-  if [ -f /opt/mempalace-toolkit/extensions/pi/mempalace.ts ] && \
-     command -v mempalace &>/dev/null && \
-     [ ! -L "$HOME/.pi/agent/extensions/mempalace.ts" ]; then
-    ln -sf /opt/mempalace-toolkit/extensions/pi/mempalace.ts \
-           "$HOME/.pi/agent/extensions/mempalace.ts"
-  fi
-
-  # pi-fork (fork tool) + pi-observational-memory (recall tool).
-  # These are pi packages (not symlink-style extensions): they're cloned to
-  # /opt with node_modules baked at BUILD time, then registered here via
-  # `pi install <local-path>`. Verified 2026-06-03: a local-path install is
-  # instant + in-place (pi loads the extension directly from /opt) + idempotent
-  # (no duplicate package entry on re-run), and stores a relative path that
-  # resolves into the image-layer /opt so it survives volume recreate. The
-  # fork/recall tools register on the NEXT pi start (extensions bind at
-  # startup). Guard on settings.json so we only install once per volume.
-  for _pkg in /opt/pi-fork /opt/pi-observational-memory; do
-    [ -d "$_pkg" ] || continue
-    _name=$(basename "$_pkg")
-    if ! grep -q "$_name" "$HOME/.pi/agent/settings.json" 2>/dev/null; then
-      pi install "$_pkg" >/dev/null 2>&1 || \
-        echo "WARN: pi install $_name failed (continuing)"
-    fi
-  done
-fi
-
 # ── Skillset: deploy skills/instructions from mounted skillset repo ──
 # When the skillset repo is mounted (at $HOME/skillset or /workspace/skillset),
 # run the deploy script to create relative symlinks for skills and instructions.
@@ -148,6 +138,116 @@ if [ -n "$SKILLSET_DEPLOY" ]; then
   "$SKILLSET_DEPLOY" --bootstrap --prune-stale >/dev/null 2>&1 || true
 fi
 
+# ── OMOS bundled skills: symlink from the image into the flat skills dir ──
+# The oh-my-opencode-slim package bundles its skills at a fixed, image-internal
+# path (npm global prefix /usr — see Dockerfile.variant). Historically the omos
+# *installer* COPIED them into ~/.config/opencode/skills/ on first run only,
+# freezing them in the persistent `devbox-opencode-config` named volume: pulling
+# a newer image never refreshed them, and the only update path was
+# `OMOS_RESET=true` (which also clobbers the user's hand-tuned opencode config).
+#
+# Instead we symlink them from the IMAGE into ~/.agents/skills/ — the same flat
+# dir skillset uses, which opencode scans (directly and via the ~/.claude/skills
+# pointer). Because the link targets live in the image, `docker compose pull` +
+# recreate updates the skills for free: no installer run, no config reset.
+#
+# ~/.agents/skills/ is authoritative. The legacy ~/.config/opencode/skills/ real
+# dir is intentionally bypassed; a one-time migration backs up (never destroys)
+# the frozen copies it holds so they stop shadowing the fresh image-sourced
+# skills. The migration marker lives in the parent config dir, not inside
+# skills/, so it never interferes with that directory's contents.
+#
+# Absolute symlink (not relative like skillset): the target is always inside the
+# container at a fixed /usr path, and ~/.agents/skills/ is an ephemeral
+# container-layer dir rebuilt each start — there is no host/container path
+# divergence to guard against. The whole block is non-fatal (`{ … } || true`):
+# a transient ln/mv failure must never brick container startup, mirroring the
+# skillset deploy above. Runs AFTER skillset deploy so OMOS wins any name
+# collision (e.g. `simplify`) via `ln -sfn`. Gated by OMOS_SKILLS (default true)
+# and the presence of the bundled skills (omos-variant images only).
+if [ "${OMOS_SKILLS:-true}" = "true" ]; then
+  OMOS_SKILLS_SRC=""
+  for cand in \
+    /usr/lib/node_modules/oh-my-opencode-slim/src/skills \
+    /usr/local/lib/node_modules/oh-my-opencode-slim/src/skills; do
+    if [ -d "$cand" ]; then OMOS_SKILLS_SRC="$cand"; break; fi
+  done
+  if [ -n "$OMOS_SKILLS_SRC" ]; then
+    {
+      AGENTS_SKILLS_DIR="$HOME/.agents/skills"
+      OPENCODE_SKILLS_DIR="$HOME/.config/opencode/skills"
+      OMOS_SKILLS_MARKER="$HOME/.config/opencode/.omos-skills-migrated"
+      mkdir -p "$AGENTS_SKILLS_DIR"
+      for skill_path in "$OMOS_SKILLS_SRC"/*/; do
+        [ -d "$skill_path" ] || continue
+        name="$(basename "$skill_path")"
+        # OMOS wins collisions: -f replaces an existing symlink (e.g. skillset's).
+        ln -sfn "${skill_path%/}" "$AGENTS_SKILLS_DIR/$name"
+        # One-time unshadow: back up — never destroy — the frozen real copy the
+        # old installer left in the persistent config volume. `! -L` ensures we
+        # only ever touch a real dir, never a symlink a user/skillset created.
+        if [ ! -f "$OMOS_SKILLS_MARKER" ] \
+          && [ -d "$OPENCODE_SKILLS_DIR/$name" ] \
+          && [ ! -L "$OPENCODE_SKILLS_DIR/$name" ]; then
+          mv "${OPENCODE_SKILLS_DIR:?}/$name" \
+            "${OPENCODE_SKILLS_DIR}/${name}.bak.$(date +%s)"
+        fi
+      done
+      touch "$OMOS_SKILLS_MARKER" 2>/dev/null || true
+    } || true
+  fi
+fi
+
+# ── Image-baked fallback skills + harness instruction ────────────
+# Baked under /usr/local/share/opencode-devbox/ (see that dir's VENDORED.md).
+# Linked into place on every start so a container behaves correctly even with
+# NO skillset mounted. Targets live in the image, so `docker compose pull` +
+# recreate refreshes them for free. Whole block is non-fatal (`{ … } || true`):
+# a transient link failure must never brick container startup.
+#
+# Precedence is intentional: this runs AFTER the skillset deploy and the OMOS
+# bundled-skills block above, and the skill links are ONLY-WHEN-ABSENT, so a
+# mounted skillset or an OMOS-bundled skill of the same name always wins; the
+# baked copies merely fill gaps.
+DEVBOX_SKILLS_SRC=/usr/local/share/opencode-devbox/skills
+if [ -d "$DEVBOX_SKILLS_SRC" ]; then
+  {
+    mkdir -p "$HOME/.agents/skills"
+    for _sk in "$DEVBOX_SKILLS_SRC"/*/; do
+      [ -d "$_sk" ] || continue
+      _skname=$(basename "$_sk")
+      # Only-when-absent: never clobber a skillset/OMOS/user skill of this name.
+      if [ ! -e "$HOME/.agents/skills/$_skname" ]; then
+        ln -s "${_sk%/}" "$HOME/.agents/skills/$_skname"
+      fi
+    done
+  } || true
+fi
+
+# Harness instructions: opencode auto-loads ~/.config/opencode/instructions/*.md
+# as a session-start system prompt. ~/.config/opencode is the persistent
+# devbox-opencode-config volume, so we SYMLINK to the fixed image path (never
+# copy — a copy would freeze in the volume) and refresh our own symlink each
+# start. A real file of the same name (e.g. one a user dropped) is left alone;
+# the baked filename (opencode-devbox.md) is distinct from skillset's
+# instructions, so there is no collision with a mounted skillset.
+DEVBOX_INSTR_SRC=/usr/local/share/opencode-devbox/instructions
+if [ -d "$DEVBOX_INSTR_SRC" ]; then
+  {
+    mkdir -p "$HOME/.config/opencode/instructions"
+    for _instr in "$DEVBOX_INSTR_SRC"/*.md; do
+      [ -f "$_instr" ] || continue
+      _iname=$(basename "$_instr")
+      _dst="$HOME/.config/opencode/instructions/$_iname"
+      # Leave a real file alone; (re)create our own symlink otherwise.
+      if [ -e "$_dst" ] && [ ! -L "$_dst" ]; then
+        continue
+      fi
+      ln -sfn "$_instr" "$_dst"
+    done
+  } || true
+fi
+
 CONFIG_DIR="$HOME/.config/opencode"
 OMOS_CONFIG="$CONFIG_DIR/oh-my-opencode-slim.json"
 
@@ -169,15 +269,14 @@ if [ "${ENABLE_OMOS:-false}" = "true" ]; then
       OMOS_TMUX_FLAG="yes"
     fi
 
-    OMOS_SKILLS_FLAG="yes"
-    if [ "${OMOS_SKILLS:-true}" = "false" ]; then
-      OMOS_SKILLS_FLAG="no"
-    fi
-
+    # Skills are NOT installer-managed any more — they are symlinked from the
+    # image into ~/.agents/skills/ by the OMOS bundled-skills block above
+    # (gated by OMOS_SKILLS). Always pass --skills=no so the installer never
+    # writes frozen copies into the persistent config volume.
     bun x oh-my-opencode-slim@latest install \
       --no-tui \
       --tmux="${OMOS_TMUX_FLAG}" \
-      --skills="${OMOS_SKILLS_FLAG}"
+      --skills=no
 
     echo "oh-my-opencode-slim configured successfully."
   else
@@ -188,13 +287,11 @@ if [ "${ENABLE_OMOS:-false}" = "true" ]; then
       echo "OMOS_RESET=true — regenerating oh-my-opencode-slim config..."
       OMOS_TMUX_FLAG="no"
       [ "${OMOS_TMUX:-false}" = "true" ] && OMOS_TMUX_FLAG="yes"
-      OMOS_SKILLS_FLAG="yes"
-      [ "${OMOS_SKILLS:-true}" = "false" ] && OMOS_SKILLS_FLAG="no"
 
       bun x oh-my-opencode-slim@latest install \
         --no-tui \
         --tmux="${OMOS_TMUX_FLAG}" \
-        --skills="${OMOS_SKILLS_FLAG}" \
+        --skills=no \
         --reset
     fi
   fi

entrypoint.sh

@@ -87,7 +87,7 @@ for dir in \
   /home/"$USER_NAME"/.vscode-server \
   /home/"$USER_NAME"/.config/opencode \
   /home/"$USER_NAME"/.config/nvim \
-  /home/"$USER_NAME"/.pi \
+  /home/"$USER_NAME"/.ssh-local \
   /home/"$USER_NAME"/.agents/skills; do
   [ -d "$dir" ] || continue
 

rootfs/etc/xdg/nvim/sysinit.vim

@@ -0,0 +1,18 @@
+" opencode-devbox — system-wide Neovim defaults.
+"
+" This is Neovim's *system vimrc*: it loads for every user before any personal
+" ~/.config/nvim, and personal configs can still override it.
+"
+" Enable 24-bit ("true") colour. Without it, Neovim's default theme is squeezed
+" into a 256-colour palette where strings/comments become a muddy, low-contrast
+" dark colour — a common complaint over ssh/kitty where COLORTERM often isn't
+" propagated into the container. Modern terminals (kitty, WezTerm, iTerm2,
+" Alacritty, ...) all support true colour; the bundled kitty-terminfo also lets
+" Neovim auto-detect it, but forcing it here guarantees readable colour
+" regardless of how the terminal type / COLORTERM reach the container.
+"
+"   Opt out for a session:  :set notermguicolors
+"   Override permanently:   set your own value in ~/.config/nvim/init.lua
+if has('termguicolors')
+  set termguicolors
+endif

rootfs/home/developer/.bash_aliases

@@ -54,6 +54,38 @@ alias gs='git status'
 alias gd='git diff'
 alias gl='git log --oneline --graph --decorate -20'
 
+# ── Host SSH reachability check (once per container lifetime) ─────────────
+# Warns at first shell startup if the Mac host is not reachable via SSH.
+# Only runs inside a container, only if the jump key exists, and only once
+# per container lifetime (/tmp flag is cleared on recreate).
+_devbox_check_host_ssh() {
+  [ -f "/.dockerenv" ] || return 0
+  local ssh_cfg="$HOME/.ssh-local/config"
+  [ -f "$ssh_cfg" ] || return 0
+  local key_pub="$HOME/.ssh-local/devbox_jump_ed25519.pub"
+  [ -f "$key_pub" ] || return 0
+  local flag="/tmp/.devbox_host_ssh_ok"
+  [ -f "$flag" ] && return 0
+  if ssh -F "$ssh_cfg" \
+        -o BatchMode=yes \
+        -o ConnectTimeout=2 \
+        -o StrictHostKeyChecking=accept-new \
+        mac true 2>/dev/null; then
+    touch "$flag"
+    return 0
+  fi
+  local pub_key
+  pub_key=$(cat "$key_pub")
+  printf '\n\033[1;33m⚠  devbox: Mac host not reachable via SSH\033[0m\n'
+  printf '   Some tools use SSH to run commands on the Mac host.\n'
+  printf '   Fix (run both on the Mac):\n\n'
+  printf '   \033[1mStep 1\033[0m  System Settings → General → Sharing → Remote Login → ON\n\n'
+  printf '   \033[1mStep 2\033[0m  echo '"'"'%s'"'"' >> ~/.ssh/authorized_keys\n' "$pub_key"
+  printf '\n   Then open a new shell in the container to verify.\n\n'
+}
+_devbox_check_host_ssh
+unset -f _devbox_check_host_ssh
+
 # ── LAN access via the host (dssh) ───────────────────────────────────
 # When running on a VM-backed host (macOS OrbStack / Docker Desktop), the
 # entrypoint's setup-lan-access.sh generates ~/.ssh-local/config so the host
@@ -89,9 +121,16 @@ fi
 # we append with a newline separator to avoid the ';;' parse error
 # described at the top of this file. Guarded so repeated sourcing
 # (e.g. `exec bash`) doesn't stack duplicates.
+#
+# The guard MUST stay shell-local (NOT exported): if it leaks into child
+# processes, every nested shell -- crucially each tmux pane, which inherits
+# the tmux server's env -- skips installing `history -a` and only persists
+# history on a clean exit. Abrupt termination (docker stop, tmux kill-server,
+# SIGKILL) then loses that shell's in-memory history. Keeping it unexported
+# means each new interactive shell re-installs its own per-prompt flush.
 if [ -z "${DEVBOX_HIST_SET:-}" ]; then
   PROMPT_COMMAND="${PROMPT_COMMAND:+$PROMPT_COMMAND$'\n'}history -a"
-  export DEVBOX_HIST_SET=1
+  DEVBOX_HIST_SET=1
 fi
 
 # ── Prompt: show [opencode-devbox] tag so it's obvious you're in the container

rootfs/home/developer/.gitignore_global

@@ -0,0 +1,15 @@
+# Global gitignore — personal/tooling artifacts (applies to all repos in the container)
+# Seeded into $HOME/.gitignore_global by entrypoint-user.sh and wired via
+# `git config --global core.excludesFile`. Edit freely; it is yours after first boot.
+
+# backup / editor / merge artifacts
+*.bak
+*.bak.*
+*~
+*.orig
+*.swp
+*.tmp
+
+# Claude Code per-machine local settings: holds machine-specific permissions
+# and can carry credentials — must never be committed.
+**/.claude/settings.local.json

rootfs/usr/local/bin/dot-watch

@@ -0,0 +1,59 @@
+#!/usr/bin/env bash
+# dot-watch — auto-rerender a graphviz .dot file to PNG on every save.
+#
+# WHY THIS EXISTS
+#   pi-studio renders mermaid natively but has no graphviz/DOT renderer.
+#   Its markdown preview DOES render local image links (.png/.jpg/.gif/.webp),
+#   and the editor offers "refresh from disk". This helper closes the loop:
+#   edit a .dot file -> dot-watch regenerates <name>.png -> hit refresh in
+#   Studio to see the update. Uses mtime polling (no inotify dependency,
+#   which isn't in the trixie-slim base).
+#
+# USAGE
+#   dot-watch <file.dot> [layout] [dpi]
+#     layout: dot|neato|fdp|circo|twopi  (default: dot)
+#     dpi:    output resolution           (default: 150)
+#   env: DOT_WATCH_INTERVAL=<seconds>     poll interval (default: 1)
+#
+# EXAMPLES
+#   dot-watch /workspace/graph.dot
+#   dot-watch graph.dot neato 200
+
+set -euo pipefail
+
+SRC="${1:?usage: dot-watch <file.dot> [layout] [dpi]}"
+LAYOUT="${2:-dot}"
+DPI="${3:-150}"
+
+[[ -f "$SRC" ]] || { echo "error: no such file: $SRC" >&2; exit 1; }
+command -v "$LAYOUT" >/dev/null || { echo "error: layout engine '$LAYOUT' not found" >&2; exit 1; }
+
+OUT="${SRC%.dot}.png"
+INTERVAL="${DOT_WATCH_INTERVAL:-1}"   # seconds between polls
+ERRLOG="$(mktemp -t dot-watch.XXXXXX.err)"
+trap 'rm -f "$ERRLOG"' EXIT
+
+render() {
+    if "$LAYOUT" -Tpng -Gdpi="$DPI" "$SRC" -o "$OUT" 2> "$ERRLOG"; then
+        printf '[%s] rendered -> %s\n' "$(date +%H:%M:%S)" "$OUT"
+    else
+        printf '[%s] DOT error:\n' "$(date +%H:%M:%S)"
+        sed 's/^/    /' "$ERRLOG"
+    fi
+}
+
+# portable mtime (GNU stat, fallback to BSD stat)
+mtime() { stat -c %Y "$1" 2>/dev/null || stat -f %m "$1" 2>/dev/null; }
+
+echo "watching $SRC ($LAYOUT, ${DPI}dpi) -> $OUT  [Ctrl-C to stop]"
+render
+last="$(mtime "$SRC")"
+while true; do
+    sleep "$INTERVAL"
+    [[ -f "$SRC" ]] || continue
+    now="$(mtime "$SRC")"
+    if [[ "$now" != "$last" ]]; then
+        last="$now"
+        render
+    fi
+done

rootfs/usr/local/bin/opencode-devbox-version

@@ -0,0 +1,88 @@
+#!/usr/bin/env bash
+# opencode-devbox-version — show which opencode-devbox image build is running.
+#
+# WHY THIS EXISTS
+#   The image bakes ground-truth build info into
+#   /etc/opencode-devbox/build-manifest.json at `docker build` time (see
+#   Dockerfile.variant): the release tag, build date, source commit, the live
+#   `opencode --version` at build time, the installed oh-my-opencode-slim
+#   version (omos variant only), and the actual checked-out commit of the
+#   /opt/mempalace-toolkit clone. That answers "what image am I running?" —
+#   but only if you know to go look for the file. This wraps it into one
+#   command, prints it human-first at container start (see entrypoint-user.sh),
+#   and stays available on demand for the rest of the session.
+#
+# USAGE
+#   opencode-devbox-version           human-readable summary (default)
+#   opencode-devbox-version --json    raw manifest JSON (for scripting)
+#   opencode-devbox-version --quiet   one-line "release_tag (source_revision)"
+#
+# EXIT STATUS
+#   0 on success. 1 if the manifest is missing (e.g. an image built before
+#   this file existed, or a non-opencode-devbox base) — prints a short notice
+#   to stderr rather than failing silently.
+
+set -euo pipefail
+
+MANIFEST=/etc/opencode-devbox/build-manifest.json
+MODE="human"
+
+case "${1:-}" in
+  --json)  MODE="json" ;;
+  --quiet|-q) MODE="quiet" ;;
+  --help|-h)
+    sed -n '2,22p' "$0" | sed 's/^# \?//'
+    exit 0
+    ;;
+esac
+
+if [ ! -f "$MANIFEST" ]; then
+  echo "opencode-devbox-version: no build manifest at $MANIFEST" >&2
+  echo "  (image predates the manifest, or this isn't an opencode-devbox image)" >&2
+  exit 1
+fi
+
+if ! command -v jq >/dev/null 2>&1; then
+  echo "opencode-devbox-version: jq not found; dumping raw manifest instead" >&2
+  cat "$MANIFEST"
+  exit 0
+fi
+
+if [ "$MODE" = "json" ]; then
+  cat "$MANIFEST"
+  exit 0
+fi
+
+release_tag=$(jq -r '.release_tag' "$MANIFEST")
+build_date=$(jq -r '.build_date' "$MANIFEST")
+source_rev=$(jq -r '.source_revision' "$MANIFEST")
+opencode_version_baked=$(jq -r '.opencode_version' "$MANIFEST")
+
+if [ "$MODE" = "quiet" ]; then
+  printf '%s (%s)\n' "$release_tag" "${source_rev:0:7}"
+  exit 0
+fi
+
+# Live drift check: has `opencode` been upgraded since this container was built?
+# The image is immutable, but `npm install -g` as the developer user lands on the
+# persistent devbox-opencode-config volume (NPM_CONFIG_PREFIX is
+# ~/.config/opencode/npm-global), which CAN shadow the baked /usr binary. So we
+# report the live version and flag a mismatch rather than trusting the manifest
+# blindly — same "ground truth over intent" spirit as how the manifest itself is
+# generated in Dockerfile.variant.
+opencode_version_live=""
+if command -v opencode >/dev/null 2>&1; then
+  opencode_version_live=$(opencode --version 2>/dev/null | head -n1 | tr -d '\r\n')
+fi
+
+printf 'opencode-devbox %s\n' "$release_tag"
+printf '  built:    %s  (source %s)\n' "$build_date" "${source_rev:0:12}"
+if [ -n "$opencode_version_live" ] && [ "$opencode_version_live" != "$opencode_version_baked" ]; then
+  printf '  opencode: %s  \033[33m(baked as %s — drift detected)\033[0m\n' \
+    "$opencode_version_live" "$opencode_version_baked"
+else
+  printf '  opencode: %s\n' "${opencode_version_live:-$opencode_version_baked}"
+fi
+
+printf '  components:\n'
+jq -r '.components | to_entries[] | select(.value != null) | "    \(.key): \(.value[0:12])"' "$MANIFEST"

rootfs/usr/local/lib/opencode-devbox/generate-config.py

@@ -3,9 +3,12 @@
 Generate opencode.json from environment variables on first container start.
 
 Safety guarantees:
-  - NEVER overwrites an existing opencode.json. If the file is present
-    (whether bind-mounted from the host, persisted in a named volume, or
-    previously generated), this script exits immediately without writing.
+  - NEVER overwrites an existing config (opencode.json / opencode.jsonc),
+    whether bind-mounted from the host, persisted in a named volume, or
+    previously generated. When a config already exists, this script instead
+    writes a NON-loaded `opencode.jsonc.proposed` sidecar (only when the
+    freshly-generated config would differ) so new image defaults can be
+    reviewed and merged by hand. opencode never loads the .proposed file.
   - Requires OPENCODE_PROVIDER to be set. Without it, no file is written.
 
 Environment variables:
@@ -13,29 +16,43 @@ Environment variables:
   OPENCODE_MODEL     Optional. Overrides the provider default model.
   AWS_REGION         Bedrock only. Default: us-east-1.
   AWS_PROFILE        Bedrock only. Default: default.
+  MEMPALACE_REMOTE_URL    Optional. Point MemPalace at a shared HTTP endpoint
+                          instead of spawning a local stdio server.
+  MEMPALACE_REMOTE_TOKEN  Optional. Bearer token for MEMPALACE_REMOTE_URL.
 
 MCP servers are auto-registered for tools detected on PATH:
-  - mempalace (if installed) — enabled
+  - mempalace — LOCAL stdio if `mempalace-mcp` is installed, or EXTERNAL
+    (remote HTTP) when MEMPALACE_REMOTE_URL is set (shared palace)
   - gitea-mcp (if installed) — registered but disabled by default
 
-Output path: $HOME/.config/opencode/opencode.json
+Output path: $HOME/.config/opencode/opencode.jsonc
+             (existing config preserved; newer defaults surfaced as
+              $HOME/.config/opencode/opencode.jsonc.proposed)
 """
 
 from __future__ import annotations
 
 import json
 import os
+import re
 import shutil
 import sys
 from pathlib import Path
 
-# Default model per provider. Update here when upstream changes.
+# Default model per provider. Update here when upstream ships a newer model.
+#
+# THESE ARE ONLY DEFAULTS — they apply when OPENCODE_MODEL is unset, and only
+# for the provider selected by OPENCODE_PROVIDER. Any of them is overridden by
+# setting OPENCODE_MODEL=<provider>/<model> in .env (see .env.example), and the
+# generated opencode.jsonc is never overwritten on later starts, so a hand-edit
+# of the config also survives. Users who don't use Anthropic or Bedrock should
+# set OPENCODE_PROVIDER (and OPENCODE_MODEL if the per-provider default below
+# isn't what they want) rather than editing this file — see the README section
+# "Choosing a provider and model".
 DEFAULT_MODELS: dict[str, str] = {
-    "anthropic": "anthropic/claude-sonnet-4-6",
-    "openai": "openai/gpt-5.4",
-    "amazon-bedrock": (
-        "amazon-bedrock/global.anthropic.claude-sonnet-4-5-20250929-v1:0"
-    ),
+    "anthropic": "anthropic/claude-opus-5",
+    "openai": "openai/gpt-5.6",
+    "amazon-bedrock": "amazon-bedrock/global.anthropic.claude-opus-5",
 }
 
 # Fallback when OPENCODE_PROVIDER is set but not recognized.
@@ -74,13 +91,27 @@ def register_mcp_servers(config: dict) -> list[str]:
     """
     servers: dict[str, dict] = {}
 
-    # MemPalace — local-first AI memory (if installed).
-    # `mempalace-mcp` is the entry-point binary shipped by the mempalace
-    # Python package. `uv tool install mempalace` places it on PATH as a
-    # shim whose shebang points at the isolated venv's Python, so system
-    # `python3 -m mempalace.mcp_server` (which would fail — system
-    # python3 can't import from the uv venv) is unnecessary here.
-    if shutil.which("mempalace-mcp"):
+    # MemPalace — AI memory. Two modes, same env contract as the mempalace.ts
+    # pi extension (mempalace-toolkit), so one shared MemPalace can serve pi +
+    # opencode + native:
+    #   LOCAL (default): stdio subprocess. `mempalace-mcp` is the entry-point
+    #     binary shipped by the mempalace Python package. `uv tool install
+    #     mempalace` places it on PATH as a shim whose shebang points at the
+    #     isolated venv's Python, so system `python3 -m mempalace.mcp_server`
+    #     (which would fail — system python3 can't import from the uv venv) is
+    #     unnecessary here.
+    #   EXTERNAL: set MEMPALACE_REMOTE_URL to a shared MemPalace HTTP endpoint
+    #     (e.g. http://mempalace.lan:8765/mcp). MEMPALACE_REMOTE_TOKEN, if set,
+    #     becomes a Bearer auth header. No local mempalace-mcp is spawned, so
+    #     the binary need not even be installed in this mode.
+    remote_url = os.environ.get("MEMPALACE_REMOTE_URL", "").strip()
+    if remote_url:
+        entry: dict = {"type": "remote", "url": remote_url}
+        token = os.environ.get("MEMPALACE_REMOTE_TOKEN", "").strip()
+        if token:
+            entry["headers"] = {"Authorization": f"Bearer {token}"}
+        servers["mempalace"] = entry
+    elif shutil.which("mempalace-mcp"):
         servers["mempalace"] = {
             "type": "local",
             "command": ["mempalace-mcp"],
@@ -110,6 +141,113 @@ def register_mcp_servers(config: dict) -> list[str]:
     return list(servers.keys())
 
 
+def render_config(provider: str, model: str) -> tuple[dict, str, list[str]]:
+    """Build the config dict and its JSONC rendering for a provider/model.
+
+    Shared by first-generation and the proposed-config side-channel so the
+    two can never drift. Returns (config_dict, jsonc_text, mcp_servers_added).
+    """
+    config = build_config(provider, model)
+    added = register_mcp_servers(config)
+
+    # Write as JSONC so we can include helpful comments.
+    content = json.dumps(config, indent=2)
+    # Insert a comment about the Context7 API key after the context7 url line.
+    context7_comment = (
+        '      "url": "https://mcp.context7.com/mcp"\n'
+        "      // For higher rate limits, sign up at https://context7.com/dashboard\n"
+        '      // and add: "headers": { "CONTEXT7_API_KEY": "{env:CONTEXT7_API_KEY}" }'
+    )
+    content = content.replace(
+        '      "url": "https://mcp.context7.com/mcp"',
+        context7_comment,
+    )
+    return config, content, added
+
+
+def _loads_jsonc(text: str) -> dict:
+    """Parse JSONC (JSON + // line comments), preserving // inside strings.
+
+    Uses the same string-aware comment stripper as scripts/smoke-test.sh, so a
+    value such as an https:// URL is never corrupted. Raises on invalid JSON
+    (e.g. trailing commas) — callers treat that as 'cannot compare'.
+    """
+    pattern = r'"(?:\\.|[^"\\])*"|//[^\n]*'
+    stripped = re.sub(
+        pattern,
+        lambda m: m.group(0) if m.group(0).startswith('"') else "",
+        text,
+    )
+    return json.loads(stripped)
+
+
+PROPOSED_HEADER = """\
+// ───────────────────────────────────────────────────────────────
+// PROPOSED opencode config — NOT loaded by opencode.
+//
+// This is what opencode-devbox would generate for your CURRENT environment
+// plus THIS image's defaults. It is written only when it differs from your
+// live opencode.jsonc, as a manual-merge reference — e.g. a newer image added
+// a default MCP server you do not have yet. opencode only loads
+// opencode.json / opencode.jsonc, never this .proposed file.
+//
+// NOTE: this reflects env + image defaults, so a difference may be a new image
+// default OR simply one of your own past edits (changed model, gitea
+// enabled=true, …). Diff against your live config and merge what you want.
+// Delete this file any time — it is rewritten on the next start if still
+// relevant, and removed automatically once your live config matches.
+// ───────────────────────────────────────────────────────────────
+"""
+
+
+def write_proposed(
+    proposed_file: Path, live_file: Path, config: dict, content: str
+) -> None:
+    """Non-destructively surface a newer default config beside the live one.
+
+    Writes <proposed_file> ONLY when the freshly-rendered config differs from
+    the live config (or the live config cannot be parsed for comparison).
+    Removes a stale proposed file when the live config already matches. NEVER
+    touches the live config itself.
+    """
+    try:
+        live = _loads_jsonc(live_file.read_text())
+        differs = live != config
+        comparable = True
+    except (OSError, ValueError):
+        # Can't read or parse the live config — surface the proposal rather
+        # than silently guess they are equivalent.
+        comparable = False
+        differs = True
+
+    if comparable and not differs:
+        if proposed_file.exists():
+            try:
+                proposed_file.unlink()
+                print(
+                    f"Live opencode config matches image defaults; removed "
+                    f"stale {proposed_file.name}.",
+                    file=sys.stderr,
+                )
+            except OSError:
+                pass
+        return
+
+    try:
+        proposed_file.write_text(PROPOSED_HEADER + content + "\n")
+    except OSError as e:
+        print(f"WARN: could not write {proposed_file}: {e}", file=sys.stderr)
+        return
+
+    why = "" if comparable else " (existing config could not be parsed for comparison)"
+    print(
+        f"A newer default opencode config is available at {proposed_file}{why}. "
+        "It is NOT applied automatically — diff/merge it into your live config "
+        "manually, or delete it to dismiss.",
+        file=sys.stderr,
+    )
+
+
 def main() -> int:
     provider = os.environ.get("OPENCODE_PROVIDER", "").strip()
     if not provider:
@@ -120,19 +258,7 @@ def main() -> int:
     config_dir = home / ".config" / "opencode"
     config_file = config_dir / "opencode.jsonc"
     config_file_legacy = config_dir / "opencode.json"
-
-    # CRITICAL: never overwrite an existing config. Users may have
-    # bind-mounted their host config directory, or their config may be
-    # persisted in a named volume from a previous run.
-    # Check both .json and .jsonc variants.
-    if config_file.exists() or config_file_legacy.exists():
-        existing = config_file if config_file.exists() else config_file_legacy
-        print(
-            f"Existing config found at {existing} — "
-            "skipping generation.",
-            file=sys.stderr,
-        )
-        return 0
+    proposed_file = config_dir / "opencode.jsonc.proposed"
 
     if provider not in DEFAULT_MODELS:
         print(
@@ -145,30 +271,37 @@ def main() -> int:
         provider, FALLBACK_MODEL
     )
 
+    config, content, added = render_config(provider, model)
+
+    # CRITICAL: never overwrite an existing config. Users may have bind-mounted
+    # their host config directory, or their config may be persisted in a named
+    # volume from a previous run. When a config already exists we instead
+    # surface any newer image defaults via a NON-loaded opencode.jsonc.proposed
+    # sidecar for manual merge (see write_proposed) — the live file is untouched.
+    existing = None
+    if config_file.exists():
+        existing = config_file
+    elif config_file_legacy.exists():
+        existing = config_file_legacy
+    if existing is not None:
+        print(
+            f"Existing config found at {existing} — not overwritten.",
+            file=sys.stderr,
+        )
+        write_proposed(proposed_file, existing, config, content)
+        return 0
+
     print(f"Generating opencode config for provider: {provider}", file=sys.stderr)
-
-    config = build_config(provider, model)
-    added = register_mcp_servers(config)
-
     config_dir.mkdir(parents=True, exist_ok=True)
-
-    # Write as JSONC so we can include helpful comments.
-    content = json.dumps(config, indent=2)
-
-    # Insert a comment about Context7 API key after the context7 url line.
-    context7_comment = (
-        '      "url": "https://mcp.context7.com/mcp"\n'
-        "      // For higher rate limits, sign up at https://context7.com/dashboard\n"
-        '      // and add: "headers": { "CONTEXT7_API_KEY": "{env:CONTEXT7_API_KEY}" }'
-    )
-    content = content.replace(
-        '      "url": "https://mcp.context7.com/mcp"',
-        context7_comment,
-    )
-
     with config_file.open("w") as f:
         f.write(content)
         f.write("\n")
+    # The fresh config now equals the image defaults — clear any stale proposal.
+    if proposed_file.exists():
+        try:
+            proposed_file.unlink()
+        except OSError:
+            pass
 
     if added:
         print(

rootfs/usr/local/lib/opencode-devbox/setup-lan-access.sh

@@ -14,7 +14,9 @@
 #   The one thing reachable from a container on every OS is the host itself
 #   (host.docker.internal). So on VM-backed hosts we generate a writable SSH
 #   config that reaches the host and lets the user ProxyJump onward to LAN
-#   peers the host can reach. On native Linux we do nothing.
+#   peers the host can reach. On native Linux we render the same writable
+#   config (for the ControlPath redirect + Include ~/.ssh/config) but emit no
+#   jump block, since LAN peers are reachable directly there.
 #
 #   We ship the MECHANISM (a generic `host` jump alias + writable config),
 #   never the POLICY: the user's specific target hosts live in their own
@@ -30,13 +32,39 @@
 #
 # CONTROLS (env)
 #   DEVBOX_LAN_ACCESS = auto (default) | jump | off
-#       auto  → set up the jump config only on VM-backed hosts; no-op on Linux.
+#       auto  → set up the host jump only on VM-backed hosts. The writable
+#               sidecar config (ControlPath redirect + Include) is always
+#               rendered, on every OS.
 #       jump  → always set up (e.g. native Linux with extra_hosts host-gateway).
 #       off   → do nothing.
 #   HOST_SSH_USER  — the username to SSH into the host as. REQUIRED for the
 #       jump to authenticate. If unset we still generate the config but print
 #       a hint with the public key to authorize on the host.
 #   DEVBOX_HOST_ALIAS — host hostname to reach (default host.docker.internal).
+#   DEVBOX_LAN_AUTOJUMP_PRIVATE = 0 (default) | 1
+#       1 → also emit a catch-all that ProxyJumps *any* RFC1918 (private) IP
+#       through the host. Lets bare `dssh user@<private-IP>` work on whatever
+#       LAN the (roaming) host is currently joined to, without naming peers.
+#       Matches by the address you TYPE, not the resolved HostName, so it never
+#       overrides named hosts that already carry their own ProxyJump.
+#
+# HOST-OWNED PEER POLICY (portable; keeps this image generic)
+#   Named LAN peers are facts about a *specific* host's network, not about the
+#   image — a roaming laptop sees different LANs. So we never bake peer names
+#   here. Instead, if the host bind-mounts ~/.config/devbox-shell/ssh-lan.conf
+#   (the same devbox-shell bridge dir used for shared aliases), we Include it
+#   *before* ~/.ssh/config. That file holds the host's own jump overrides, e.g.
+#       Host pve pve-2 pbs-vm
+#           ProxyJump host
+#   First-value-wins means ProxyJump is taken from there while HostName/User/
+#   IdentityFile are inherited from the matching block in ~/.ssh/config.
+#
+# SCOPING NOTE (important)
+#   `Include` is scoped to the enclosing Host/Match block. So every Include
+#   below is preceded by a bare `Host *` to reset the active context to
+#   match-all — otherwise the included config would only apply when targeting
+#   `host`/`mac` and named peers like `pve` would silently fall back to ssh
+#   defaults.
 #
 # Idempotent: re-renders the config every run (cheap); never regenerates the
 # key. Always non-fatal — never blocks container startup.
@@ -60,44 +88,53 @@ is_vm_backed() {
   getent hosts "$HOST_ALIAS_HOSTNAME" >/dev/null 2>&1
 }
 
-if [ "$MODE" = "auto" ] && ! is_vm_backed; then
-  # Native Linux host: LAN peers are reachable directly. Nothing to do.
-  exit 0
-fi
-
-# From here: MODE=jump, or MODE=auto on a VM-backed host.
-
-command -v ssh-keygen >/dev/null 2>&1 || exit 0
-
+# ── Writable socket dir + sidecar (ALWAYS, every host OS) ─────────────
+# The ControlPath redirect in the generated config needs a writable directory
+# regardless of host OS or jump mode. ~/.ssh is typically read-only, so the
+# master socket lives under the writable ~/.ssh-local. We create it and render
+# the config UNCONDITIONALLY so the redirect (and `Include ~/.ssh/config`) works
+# even on native Linux — where we set up no host jump but a read-only ~/.ssh
+# would otherwise still break ControlMaster sockets.
 mkdir -p "${SSH_LOCAL}/cm" 2>/dev/null || true
 chmod 700 "${SSH_LOCAL}" "${SSH_LOCAL}/cm" 2>/dev/null || true
 
-# ── Jump key (generated once; preserved across restarts) ──────────────
-if [ ! -f "$KEY" ]; then
-  ssh-keygen -t ed25519 -N '' -C "devbox-jump@${HOSTNAME:-container}" -f "$KEY" >/dev/null 2>&1 || exit 0
+# ── Decide whether to set up the host jump ────────────────────────────
+# Jump = reach the container host (host.docker.internal) as an SSH ProxyJump
+# onward to the host's LAN peers. Needed on VM-backed hosts (macOS / Docker
+# Desktop) or when forced with DEVBOX_LAN_ACCESS=jump. On native Linux LAN
+# peers are reachable directly, so NEED_JUMP=0 and we emit no jump block — but
+# we still render the config for the ControlPath redirect + Include.
+NEED_JUMP=0
+if [ "$MODE" = "jump" ] || { [ "$MODE" = "auto" ] && is_vm_backed; }; then
+  NEED_JUMP=1
+fi
+
+# ── Jump key (only when a jump is needed; generated once, preserved) ──
+# Persisted via a named volume on ~/.ssh-local (see compose), so a fresh key
+# is generated only on the very first start (or if the volume is wiped). When
+# we DO generate one it must be (re-)authorized on the host, so we flag it and
+# print a copy-paste authorize line below.
+KEY_JUST_GENERATED=0
+if [ "$NEED_JUMP" = "1" ] && command -v ssh-keygen >/dev/null 2>&1 && [ ! -f "$KEY" ]; then
+  if ssh-keygen -t ed25519 -N '' -C "devbox-jump@${HOSTNAME:-container}" -f "$KEY" >/dev/null 2>&1; then
     chmod 600 "$KEY" 2>/dev/null || true
+    KEY_JUST_GENERATED=1
+  fi
 fi
 
 # ── Render the writable config ────────────────────────────────────────
-USER_LINE=""
-if [ -n "${HOST_SSH_USER:-}" ]; then
+# Jump-specific blocks (the host alias, host-owned peer overrides, and the
+# optional RFC1918 catch-all) only make sense when a jump is set up; on native
+# Linux they are all empty and only the ControlPath redirect + Include remain.
+JUMP_BLOCK=""
+LAN_CONF_BLOCK=""
+AUTOJUMP_BLOCK=""
+if [ "$NEED_JUMP" = "1" ]; then
+  USER_LINE=""
+  if [ -n "${HOST_SSH_USER:-}" ]; then
     USER_LINE="    User ${HOST_SSH_USER}"
-fi
-
-INCLUDE_LINE=""
-if [ -r "${HOME}/.ssh/config" ]; then
-  INCLUDE_LINE="Include ~/.ssh/config"
-fi
-
-cat > "$CONFIG" <<EOF
-# AUTO-GENERATED by setup-lan-access.sh on every container start. Do not edit
-# by hand — edits are overwritten. Used via: ssh -F ~/.ssh-local/config <host>
-# (or the dssh / dscp aliases). See the script header for the full rationale.
-
-# ~/.ssh is typically mounted read-only, so keep our own known_hosts here.
-Host *
-    UserKnownHostsFile ~/.ssh-local/known_hosts
-    StrictHostKeyChecking accept-new
+  fi
+  JUMP_BLOCK=$(cat <<EOF
 
 # The container host (OrbStack / Docker Desktop). 'host' and 'mac' are aliases.
 Host host mac
@@ -109,25 +146,108 @@ ${USER_LINE}
     ControlPath ~/.ssh-local/cm/%r@%h:%p
     ControlPersist 4h
     ServerAliveInterval 30
+EOF
+)
 
-# Your own target hosts: add 'ProxyJump host' to their entries in your
-# bind-mounted ~/.ssh/config, pulled in below.
-${INCLUDE_LINE}
+  # Optional host-owned named-peer jump overrides (portable: lives on the host,
+  # not in the image). Included BEFORE ~/.ssh/config so its ProxyJump wins.
+  SSH_LAN_CONF="${HOME}/.config/devbox-shell/ssh-lan.conf"
+  if [ -r "$SSH_LAN_CONF" ]; then
+    LAN_CONF_BLOCK=$(cat <<'EOF'
+
+# Host-owned named-peer jump overrides (bind-mounted; edit on the host).
+# Scope reset to match-all so the Include applies to every target host.
+Host *
+Include ~/.config/devbox-shell/ssh-lan.conf
+EOF
+)
+  fi
+
+  # Optional opt-in RFC1918 catch-all: ProxyJump every private IP through the
+  # host. Matches the typed address, never the resolved HostName, so named hosts
+  # with their own ProxyJump are unaffected. Network-agnostic → roaming-safe.
+  if [ "${DEVBOX_LAN_AUTOJUMP_PRIVATE:-0}" = "1" ]; then
+    AUTOJUMP_BLOCK=$(cat <<'EOF'
+
+# RFC1918 auto-jump (DEVBOX_LAN_AUTOJUMP_PRIVATE=1): reach any private IP on
+# the host's CURRENT LAN via bare `dssh user@<ip>`. Public IPs are unmatched
+# and go direct via the container's NAT egress. NOTE: also matches the
+# container's own bridge subnet and any private IP the host can't actually
+# reach — for non-LAN private hosts behind a different jump, use their named
+# entry (which matches first by name and keeps its own ProxyJump).
+Host 10.* 192.168.* 172.16.* 172.17.* 172.18.* 172.19.* 172.20.* 172.21.* 172.22.* 172.23.* 172.24.* 172.25.* 172.26.* 172.27.* 172.28.* 172.29.* 172.30.* 172.31.*
+    ProxyJump host
+EOF
+)
+  fi
+fi
+
+INCLUDE_BLOCK=""
+if [ -r "${HOME}/.ssh/config" ]; then
+  INCLUDE_BLOCK=$(cat <<'EOF'
+
+# Your own target hosts. Scope reset to match-all so this Include applies to
+# every target (an Include is otherwise scoped to the enclosing Host block).
+# To make a LAN peer jump via the host, add 'ProxyJump host' to its entry in
+# the host-owned ~/.config/devbox-shell/ssh-lan.conf (Included above) — NOT
+# here in ~/.ssh/config, which is typically bind-mounted read-only.
+Host *
+Include ~/.ssh/config
+EOF
+)
+fi
+
+cat > "$CONFIG" <<EOF
+# AUTO-GENERATED by setup-lan-access.sh on every container start. Do not edit
+# by hand — edits are overwritten. Used via: ssh -F ~/.ssh-local/config <host>
+# (or the dssh / dscp aliases). See the script header for the full rationale.
+
+# ~/.ssh is typically mounted read-only, so keep our own known_hosts here.
+# Also redirect ControlPath into the writable sidecar: the bind-mounted
+# ~/.ssh/config commonly sets 'ControlPath ~/.ssh/cm/...' for CGNAT multiplexing,
+# but ~/.ssh is read-only here so the master socket can't be created and those
+# hosts fail to connect. First-value-wins: setting it here (before the Include)
+# overrides the read-only path for every host. Harmless when ControlMaster is off.
+Host *
+    UserKnownHostsFile ~/.ssh-local/known_hosts
+    StrictHostKeyChecking accept-new
+    ControlPath ~/.ssh-local/cm/%r@%h:%p
+${JUMP_BLOCK}
+${LAN_CONF_BLOCK}
+${AUTOJUMP_BLOCK}
+${INCLUDE_BLOCK}
 EOF
 chmod 600 "$CONFIG" 2>/dev/null || true
 
-# ── One-time hint when we can't authenticate yet ──────────────────────
+# ── Authorize hints ───────────────────────────────────────────────────
+# Print the copy-paste authorize line whenever we either (a) can't yet
+# authenticate (HOST_SSH_USER unset) or (b) just generated a NEW key that the
+# host won't recognize. With ~/.ssh-local persisted via a named volume, case
+# (b) fires only on first-ever start (or after the volume is reset) — so this
+# is normally a one-time, one-line step per machine, with no file to locate.
+if [ "$NEED_JUMP" = "1" ]; then
+PUBKEY_TEXT="$(cat "${KEY}.pub" 2>/dev/null)"
 if [ -z "${HOST_SSH_USER:-}" ]; then
   cat <<EOF
 [devbox] LAN-access jump config generated at ~/.ssh-local/config, but
          HOST_SSH_USER is unset so it can't authenticate to the host yet.
          To enable container -> host -> LAN-peer access:
            1. Set HOST_SSH_USER=<your host username> in the container env.
-           2. Authorize this key on the host (append to ~/.ssh/authorized_keys):
-                $(cat "${KEY}.pub" 2>/dev/null)
+           2. Authorize this key on the host (run ON THE HOST, once):
+                echo '${PUBKEY_TEXT}' >> ~/.ssh/authorized_keys
            3. Ensure the host's SSH server (Remote Login) is enabled.
          Then: dssh host   (or add 'ProxyJump host' to targets in ~/.ssh/config)
 EOF
+elif [ "$KEY_JUST_GENERATED" = "1" ]; then
+  cat <<EOF
+[devbox] Generated a NEW LAN-jump key. Authorize it on the host (${HOST_SSH_USER}@host),
+         then 'dssh host' and your LAN peers will work. Run this ONCE, ON THE HOST:
+           echo '${PUBKEY_TEXT}' >> ~/.ssh/authorized_keys
+         (Ensure the host's SSH server / Remote Login is enabled.)
+         This key is persisted in the ~/.ssh-local volume, so you won't need to
+         repeat this on container updates — only if that volume is reset.
+EOF
+fi
 fi
 
 exit 0

rootfs/usr/local/share/opencode-devbox/instructions/opencode-devbox.md

@@ -0,0 +1,37 @@
+<!-- opencode-devbox:managed-instruction — baked into the image at
+     /usr/local/share/opencode-devbox/instructions/opencode-devbox.md and
+     symlinked into ~/.config/opencode/instructions/ by entrypoint-user.sh on
+     every container start (opencode auto-loads instructions/*.md as a
+     session-start system prompt). A mounted skillset deploys its own
+     instructions under different filenames, so this never collides. Keep this
+     short: it is a pointer, the depth lives in the skills. -->
+
+## Running inside opencode-devbox
+
+If the directory `/usr/local/lib/opencode-devbox/` exists (or your shell prompt
+is prefixed `[devbox]`, or `~/.ssh-local/config` is present), you are in an
+**opencode-devbox container** — a Docker environment whose persistence,
+networking, DNS, host/LAN reachability, and Python/REPL behaviour differ from a
+normal workstation. Before any task that touches **reaching the host or its
+LAN, SSH, DNS/name resolution, what survives container recreate, running
+Python/REPLs, or the oh-my-opencode-slim variant**, read
+`~/.agents/skills/opencode-devbox-environment/SKILL.md`.
+
+Key reflex from that skill: **the deployment specifics are not universal** — the
+host OS, hostnames, internal domains, and nameservers vary per instance and must
+be discovered at runtime, never assumed. And interactive shell aliases
+(`dssh`, `dscp`, `cat`→`bat`) do **not** exist in your non-interactive bash
+tool, so spell out the underlying command (e.g.
+`ssh -F "$HOME/.ssh-local/config" mac …`).
+
+## Session start: load the mempalace skill
+
+If MemPalace MCP tools (e.g. `mempalace_status`, `mempalace_search`,
+`mempalace_diary_write`) are in your tool list, **read
+`~/.agents/skills/mempalace/SKILL.md` before doing non-trivial work** and follow
+its protocol: search the palace before answering about past work, and write a
+diary entry before the session ends. This is especially load-bearing here — an
+opencode-devbox container is frequently recreated, so the palace is your only
+memory across recreates. Without the habit it is just storage, not memory.
+(The skill is the consumer side; feeding the palace is the separate
+`opencode-mempalace-bridge` skill, if present.)

rootfs/usr/local/share/opencode-devbox/skills/VENDORED.md

@@ -0,0 +1,47 @@
+# Vendored fallback skills + harness instruction
+
+This tree is **image-baked** content that `entrypoint-user.sh` links into place
+on container start so the container behaves correctly **even when no private
+`skillset` repo is mounted**:
+
+- `skills/*` → symlinked into `~/.agents/skills/` (only when a skill of the same
+  name is not already present, so a mounted `skillset` or the OMOS bundled
+  skills always win).
+- `instructions/*.md` → symlinked into `~/.config/opencode/instructions/`, which
+  opencode auto-loads as a session-start system prompt (only when a real file of
+  the same name is not already there).
+
+| item | owner | how it gets here |
+|------|-------|------------------|
+| `skills/opencode-devbox-environment` | opencode-devbox (this repo) | authored here; the canonical copy |
+| `skills/mempalace` | the `skillset` repo | **vendored fallback** (snapshot only) |
+| `instructions/opencode-devbox.md` | opencode-devbox (this repo) | authored here; the proactive-load pointer |
+
+## Why fallbacks exist
+
+opencode discovers skills under `~/.agents/skills/` and loads harness
+instructions from `~/.config/opencode/instructions/*.md` at session start. A
+container started **without** the private `skillset` repo mounted would have
+neither — so the agent would not know the container-shaped facts and would not
+follow the MemPalace continuity protocol. Baking `opencode-devbox-environment`
+and `mempalace` closes that *availability* gap, and the baked
+`instructions/opencode-devbox.md` adds the matching *proactive-load* directive
+(read those two skills at session start) so a fresh container actually picks
+them up rather than relying on description-matching.
+
+Note there is **no `pi-extensions` skill here**: opencode has no `fork`/`recall`
+extensions (that is a pi-only concern), so the pi-devbox vendored set does not
+carry over 1:1.
+
+## Filename discipline
+
+`instructions/opencode-devbox.md` deliberately uses a name distinct from the
+skillset's `instructions/mempalace.md`. Both can be deployed at once (a mounted
+skillset adds its own); the distinct name means our fallback is collision-free
+and the entrypoint's never-overwrite-a-real-file guard never has to arbitrate.
+
+## Refreshing the snapshots
+
+    cp <skillset>/skills/mempalace/SKILL.md  skills/mempalace/SKILL.md
+
+Snapshot provenance at last refresh: skillset `63f3bf5`.

rootfs/usr/local/share/opencode-devbox/skills/mempalace/SKILL.md

@@ -0,0 +1,326 @@
+---
+name: mempalace
+description: MemPalace agent memory protocol. Use on every session to maintain continuity across conversations — search before answering about past work, write diary entries before session ends, and mine new projects into the palace. Load this skill at session start.
+---
+
+# MemPalace Agent Memory Protocol
+
+## Overview
+
+MemPalace gives you persistent memory across sessions via an MCP server. It stores project knowledge (mined from files), conversation summaries (diary entries), and entity relationships (knowledge graph). Without this protocol, you have tools but no habits — and memory without habits is just storage.
+
+**Core principle:** Storage is not memory. Storage + protocol = memory.
+
+## When to Load This Skill
+
+- At the **start of every session** (proactively, before the user asks)
+- When the user mentions **past conversations, decisions, or work**
+- When working on a **new project or repository** for the first time
+- When the user asks about **people, projects, or relationships**
+
+## Session Lifecycle
+
+### Phase 1: Wake Up (session start)
+
+Run these immediately when a session begins, before responding to the user:
+
+1. **Load palace overview:**
+   ```
+   mempalace_status
+   ```
+   This returns wing/room counts, the AAAK spec, and the memory protocol reminder.
+
+2. **Read your recent diary:**
+   ```
+   mempalace_diary_read(agent_name="<your_agent_name>", last_n=5)
+   ```
+   Scan for context about recent sessions — what was worked on, what matters, what's pending.
+
+3. **Check the knowledge graph** for the user or active project if relevant:
+   ```
+   mempalace_kg_query(entity="<project_or_person>")
+   ```
+
+Do NOT announce this to the user. Just do it silently to orient yourself.
+
+### Temporal grounding — compute time deltas, don't guess
+
+Diary entries and drawers carry real timestamps (`timestamp`, `created_at`).
+Before describing *when* something happened — "yesterday", "earlier today",
+"last week", "a while back" — **establish the current date/time first and
+compute the delta against the actual timestamp.** Get "now" from the injected
+session date or by running `date` in a shell; never infer it.
+
+**A container recreate or a fresh session is NOT a day boundary.** A devbox
+container (pi-devbox or opencode-devbox) is frequently restarted — often several
+times within the *same* day — and each restart begins a new session with a fresh
+wake-up. Do not reason "new session ⇒ last session was yesterday": two diary
+entries 90 minutes apart can straddle a container recreate. The only
+authoritative clock is the timestamp on the memory, not the session/container
+boundary.
+
+**Practical rule:** prefer explicit, checkable phrasing — e.g. "earlier today,
+~8h ago (both 2026-06-25)" — over a vague relative term. If you catch yourself
+about to write "yesterday" / "last week", subtract `now − entry.timestamp` and
+state the computed result. (Remember timestamps may be UTC while the wall clock
+is local — reconcile the offset before computing the delta.) Note too that
+session feeders can lag up to a week (see *Multi-harness palace*), so a recent
+absence in `wing_conversations` is not proof nothing happened.
+
+### Phase 2: Active Session (during work)
+
+#### Search Before You Speak
+
+Before answering questions about past work, decisions, people, or projects:
+
+```
+mempalace_search(query="<keywords>", wing="<project>")
+```
+
+**Never guess about facts that might be in the palace.** Wrong is worse than slow. Say "let me check" and query.
+
+#### Mine New Projects
+
+When working on a new codebase for the first time:
+
+1. Check if it's already mined:
+   ```
+   mempalace_list_wings
+   ```
+
+2. **Decide what to mine — docs first, code never (by default).**
+
+   The palace is for *context and intent*, not code recall. Code is better read from the working tree via `Read`/`Grep`/`glob` — always authoritative, never stale. Embedding source code produces thousands of low-signal drawers (e.g. `def __init__(self, ...)` across every class) that pollute search for years.
+
+   **Mine by default:**
+   - `*.md`, `*.rst`, `*.txt` — docs, READMEs, CHANGELOGs, architecture notes
+   - `AGENTS.md`, `CLAUDE.md`, `CONTRIBUTING.md`, design/decision docs — highest signal per byte
+   - `*.sh`, `Dockerfile`, `Makefile`, entrypoints — small, intent-bearing
+   - `*.yml`, `*.yaml`, `*.toml`, selective `*.json` (`docker-compose`, `pyproject`, `mkdocs.yml`, CI workflows) — skip lockfiles
+
+   **Do NOT mine by default:**
+   - `*.py`, `*.ts`, `*.tsx`, `*.js`, `*.go`, `*.rs`, `*.java`, `*.cpp`, `*.c`, `*.rb` — raw source code
+   - Test files, fixtures, generated code
+   - `node_modules/`, `.venv/`, `__pycache__/`, `.mypy_cache/`, `.pytest_cache/`, `.ruff_cache/` (the miner respects `.gitignore` but double-check)
+
+   Exception: if a code file *is* the documentation (e.g. a heavily-commented reference script, or a protocol definition), file it manually via `mempalace_add_drawer`.
+
+3. **Before mining**, inspect the repo to estimate drawer count:
+   ```bash
+   # Quick audit — what will actually get mined?
+   find <dir> -type f \
+     -not -path '*/.git/*' -not -path '*/node_modules/*' \
+     -not -path '*/.venv/*' -not -path '*/__pycache__/*' \
+     \( -name '*.md' -o -name '*.sh' -o -name '*.yml' -o -name '*.yaml' \
+        -o -name '*.toml' -o -name 'Dockerfile*' -o -name 'Makefile' \) | wc -l
+   ```
+   A docs-heavy repo should produce ~5–10 drawers per file. If a mine produces >15 drawers/file on average, code leaked in — investigate.
+
+4. Run the mine:
+   ```bash
+   mempalace init --yes <directory>
+   mempalace mine <directory> --agent <your_agent_name>
+   ```
+
+   The miner currently lacks a `--docs-only` or `--exclude-ext` flag (as of v3.3.3). Until it does, either:
+   - (a) Add a `mempalace.yaml` at the repo root with explicit include globs, OR
+   - (b) Mine everything, then surgically remove code-sourced drawers via SQL on `~/.mempalace/palace/chroma.sqlite3` (delete by `embedding_metadata.source_file LIKE '%.py'`), followed by `mempalace repair --yes`.
+
+5. If the CLI miner misses a file you *do* want (e.g., `.zsh`, an undocumented extension), file it manually:
+   ```
+   mempalace_add_drawer(wing="<project>", room="<aspect>", content="<verbatim content>", source_file="<path>")
+   ```
+
+6. After mining, reconnect to pick up the new embeddings:
+   ```
+   mempalace_reconnect
+   ```
+   If search errors occur after mining ("Error finding id"), repair the index:
+   ```bash
+   mempalace repair --yes
+   ```
+
+#### Track Facts in the Knowledge Graph
+
+When you learn new facts about people, projects, or relationships:
+
+```
+mempalace_kg_add(subject="ProjectX", predicate="uses", object="PostgreSQL")
+mempalace_kg_add(subject="Alice", predicate="owns", object="ProjectX", valid_from="2026-01-15")
+```
+
+When facts change (ended, no longer true):
+
+```
+mempalace_kg_invalidate(subject="Alice", predicate="works_at", object="OldCorp", ended="2026-03-01")
+```
+
+#### Cross-Reference with Tunnels
+
+When content in one project relates to another, create a tunnel:
+
+```
+mempalace_create_tunnel(
+  source_wing="project_api", source_room="endpoints",
+  target_wing="project_db", target_room="schema",
+  label="API endpoints map to these DB tables"
+)
+```
+
+#### Feeding opencode session history (opencode + mempalace-toolkit only)
+
+MemPalace has no upstream integration with [opencode](https://github.com/anomalyco/opencode) as of v3.3.3 — `hooks_cli.py` only supports `claude-code` and `codex` harnesses. Opencode persists every turn in a local SQLite DB at `~/.local/share/opencode/opencode.db`, but nothing moves that data into the palace automatically.
+
+On a machine with opencode + the [`mempalace-toolkit`](https://gitea.jordbo.se/joakimp/mempalace-toolkit) installed, session history is fed into `wing_conversations` via `mempalace-session` — either manually, or on a weekly systemd user timer / cron schedule shipped in `mempalace-toolkit/contrib/`. If this is missing, opencode conversations exist only in the local SQLite DB and are invisible to `mempalace_search`.
+
+**How to tell if it's set up:**
+
+```
+mempalace_list_wings
+```
+
+If `wing_conversations` exists and has a drawer count comparable to the user's opencode session count, session feeding is working. If it's empty or suspiciously small, suggest:
+
+1. Check if the toolkit is installed: `which mempalace-session`.
+2. If installed, suggest running `mempalace-session --dry-run` to preview and `mempalace-session` to file.
+3. If not installed, point the user at `gitea.jordbo.se/joakimp/mempalace-toolkit` for setup.
+
+**Don't try to paper over the gap by dumping turn-level content into the palace manually via `mempalace_add_drawer`** — that reinvents what `mempalace-session` does with normalization and dedup. Use the tool.
+
+Full routine (triggers, cadence, automation) is in the [`opencode-mempalace-bridge`](https://gitea.jordbo.se/joakimp/mempalace-toolkit) skill and the toolkit's `ARCHITECTURE.md` §5. The two skills pair: this one (`mempalace`) covers using the palace; that one (`opencode-mempalace-bridge`) covers feeding it from opencode.
+
+### Phase 3: Wind Down (session end)
+
+**Always write a diary entry before the session ends.** This is the most important habit.
+
+```
+mempalace_diary_write(
+  agent_name="<your_agent_name>",
+  entry="<AAAK compressed summary>",
+  topic="session-summary"
+)
+```
+
+#### Why still write diaries when sessions may be mined automatically?
+
+On machines running opencode + `mempalace-toolkit`, every session is mined into `wing_conversations` on a weekly (or user-defined) schedule. A common and incorrect conclusion: *"since every turn is captured automatically, writing a diary entry is redundant."* It isn't.
+
+Session mining captures **what was said** (every turn, verbatim). A diary captures **what the session meant** — editorial judgment by the agent who lived it:
+
+- Lessons learned, patterns noticed, pending items rolled forward
+- Meta-observations that were never said aloud during the session
+- Aggregate counts (commits shipped, bugs fixed, hours spent)
+- A compressed, recency-scannable summary for the *next* agent's wake-up
+
+Mining raw turns cannot surface these because the words don't exist verbatim — they're the agent's reflection at wind-down. Think of the split as *release notes* (diary) vs. *git log with diffs* (session mine): a repo keeps both because they answer different questions. So does the palace.
+
+**Practical rule:** automated mining does not replace Phase 3. Both systems cover each other's failure modes — a skipped diary is recovered from the raw turns; a missed mine is recovered from the diary summary. For the full treatment (comparison table, retrieval patterns, token economics), see [`mempalace-toolkit/ARCHITECTURE.md` §5 → "Diary vs session mine: why keep both?"](https://gitea.jordbo.se/joakimp/mempalace-toolkit/src/branch/main/ARCHITECTURE.md#diary-vs-session-mine-why-keep-both).
+
+#### AAAK Diary Format
+
+Write diary entries in compressed AAAK format for efficiency. Structure:
+
+```
+SESSION:<date>|<what.you.worked.on>|
+TASKS:
+1.<task.description>→<outcome>|
+2.<task.description>→<outcome>|
+DISCOVERED:<unexpected.findings>|
+ENTITIES:<people.or.projects.encountered>|
+<importance: one to five stars>
+```
+
+Example:
+
+```
+SESSION:2026-04-28|api.refactor+db.migration|
+TASKS:
+1.refactored.auth.endpoints→split.into.3.modules|
+2.added.user.roles.migration→postgres.enum.type|
+DISCOVERED:legacy.session.table.unused.since.v2|
+ENTITIES:ProjectX;Alice(reviewer)|
+***
+```
+
+Rules:
+- Use dots instead of spaces within phrases
+- Use pipes as field separators
+- Use arrows for cause/effect or transitions
+- Stars indicate session importance (one to five)
+- Keep it tight — a future agent should get the gist in seconds
+
+#### What to Capture
+
+Prioritize recording:
+- **Decisions made** and their rationale
+- **Discoveries** — things that surprised you or that a future session needs to know
+- **Unfinished work** — what's pending, what was deferred
+- **User preferences** observed during the session
+- **Entities encountered** — people, projects, tools, services
+
+### Phase 4: Fact Updates
+
+If facts changed during the session, update the knowledge graph before writing the diary:
+
+```
+mempalace_kg_invalidate(subject="...", predicate="...", object="...", ended="<today>")
+mempalace_kg_add(subject="...", predicate="...", object="...", valid_from="<today>")
+```
+
+## Palace Structure
+
+### Wings
+
+Wings are top-level categories, typically one per project or domain:
+- Named after the project directory (e.g., `cli_utils`, `opencode_devbox`)
+- Agent diaries live in `wing_<agent_name>` (e.g., `wing_orchestrator`, `wing_pi`)
+
+#### Multi-harness palace
+
+A single palace can be fed by multiple coding-agent harnesses. On this machine the palace is shared between **opencode** and **pi** (Mario Zechner's pi-coding-agent). Implications:
+
+- **`wing_conversations` mixes sources.** Both harnesses' session feeders write into the same wing. To tell them apart, look at the `source_file` metadata on each drawer:
+  - `pi_<uuid>.jsonl` → pi session
+  - `<slug>_ses_<id>.jsonl` → opencode session
+  - The first chunk of each session also carries a `| source: opencode` or `| source: pi` marker in the synthetic header line.
+- **Other wings may belong to other harnesses.** For example `wing_pi` is pi's diary, not opencode's. Don't assume every diary entry was written by you — check `agent_name` on the entry.
+- **Session feeders run on different schedules.** Pi sessions are fed Tue 03:00, opencode sessions Mon 03:00. Recent sessions from either harness can lag the palace by up to a week, so absence-of-evidence in `wing_conversations` is not evidence-of-absence for recent work.
+- **Reading another harness's diary is useful.** When orienting after a gap, `mempalace_diary_read agent_name=pi` (or whichever sibling agent has been active) often gives a fresher picture than waiting for the conversations feeder to catch up.
+
+### Rooms
+
+Rooms are aspects within a wing:
+- `fzf`, `scripts`, `configuration`, `general` — whatever the miner detects
+- Diary entries go into rooms by topic tag
+
+### Drawers
+
+Drawers hold verbatim content — never summarized, always searchable.
+
+### Tunnels
+
+Cross-wing connections linking related content across projects.
+
+### Knowledge Graph
+
+Entity-relationship triples with temporal validity. Query with `mempalace_kg_query`, browse with `mempalace_kg_timeline`.
+
+## Troubleshooting
+
+| Problem | Fix |
+|---|---|
+| "No palace found" | Run `mempalace init <dir>` then `mempalace mine <dir>` |
+| "Error finding id" after mining | Run `mempalace repair --yes` then `mempalace_reconnect` |
+| Search returns irrelevant results | Use `max_distance=1.0` for stricter matching; add `wing` filter |
+| Miner skips file types | File manually with `mempalace_add_drawer` or use `--no-gitignore` |
+| Stale results after external changes | Call `mempalace_reconnect` |
+
+## Anti-Patterns
+
+- **Don't guess when you can search.** If a question touches past work, search first.
+- **Don't infer elapsed time from session or container boundaries.** A restart isn't a new day. Compare the actual timestamp (`timestamp` / `created_at`) against the current date/time before saying "yesterday", "last week", etc.
+- **Don't skip the diary.** A session without a diary entry is a session forgotten.
+- **Don't summarize drawer content.** File verbatim — the embedding model needs the original words.
+- **Don't mine .git directories or node_modules.** The CLI miner respects .gitignore by default.
+- **Don't create duplicate drawers.** Use `mempalace_check_duplicate` before adding manually.
+- **Don't treat the palace as a task list.** It's for knowledge and context, not todos.

rootfs/usr/local/share/opencode-devbox/skills/opencode-devbox-environment/SKILL.md

@@ -0,0 +1,226 @@
+---
+name: opencode-devbox-environment
+description: >-
+  Operate correctly inside an opencode-devbox container. Load when running
+  inside opencode-devbox (detection: the directory `/usr/local/lib/opencode-devbox/`
+  exists, the shell prompt is prefixed `[devbox]`, or `~/.ssh-local/config` is
+  present) and the task touches any of: reaching the Docker host or its LAN,
+  SSH, DNS name resolution, what survives container recreate (persistence vs
+  ephemerality), running Python or other REPLs, or the oh-my-opencode-slim
+  (OMOS) variant. Covers the persistence model (and why image-owned content
+  must live under /usr, not a home dir a named volume shadows), the
+  interactive-vs-tool-shell alias gotcha (dssh/dscp/cat=bat exist only in
+  interactive bash), host + LAN SSH reachability and ControlMaster, split-horizon
+  DNS mechanisms, and uv-first Python. This skill teaches MECHANISMS only —
+  concrete hostnames, usernames, internal domains, nameservers, and even the
+  host OS vary per deployment and MUST be discovered at runtime, never assumed
+  or hardcoded.
+---
+
+# opencode-devbox environment
+
+You are (or may be) running inside **opencode-devbox**: a Docker container that
+ships opencode, MemPalace, and a curated tool stack, with the host source tree
+mounted at `/workspace`. This skill is about the *container-shaped* facts that
+change how you should act — things that are easy to get wrong because they
+differ from a normal workstation shell.
+
+> **Golden rule: this environment is a template, not a fixed deployment.**
+> The host could be macOS, Windows, or Linux. There may or may not be LAN
+> peers, a VPN, split-DNS, a skillset mount, or the `-omos` variant. Detect
+> and verify the specifics live (commands below) — do **not** assume any
+> particular hostname, domain, nameserver, or OS. Where this skill shows
+> example values they are illustrative placeholders.
+
+## 0. Am I in opencode-devbox, and what's true *here*?
+
+Cheap detection signals (any one is sufficient):
+
+```sh
+[ -d /usr/local/lib/opencode-devbox ] && echo "opencode-devbox image"
+[ -r "$HOME/.ssh-local/config" ] && echo "LAN/host SSH sidecar present"
+case "$PS1" in *'[devbox]'*) echo "interactive devbox shell";; esac
+```
+
+Then orient before acting:
+
+```sh
+cat /etc/os-release | head -2          # container distro (usually Debian)
+ls -la /usr/local/lib/opencode-devbox/ # which devbox helpers exist
+sed -n '/^Host /,$p' ~/.ssh-local/config 2>/dev/null   # host/LAN reachability, if any
+mount | grep -E ' /workspace | /home/\S+/\.ssh '       # what's bind-mounted
+```
+
+## 1. Persistence vs ephemerality — know before you write
+
+The container has **three storage tiers with very different lifetimes**. Pick
+the right one or work is silently lost on the next recreate/update.
+
+| Tier | Examples | Survives `down`? | Survives `down -v`? | Survives image update / `--force-recreate`? |
+|---|---|---|---|---|
+| **Host bind-mount** | `/workspace`, usually `~/.ssh` (ro), optionally `~/.mempalace` | yes | yes (lives on host) | yes |
+| **Named volume** | `~/.config/opencode`, `~/.local/share/opencode`, `~/.local/state/opencode`, `~/.ssh-local`, `~/.cache/bash`, `~/.local/share/{uv,nvim,zoxide}` | yes | **no** | yes |
+| **Writable container layer** | anything else: `sudo apt install …`, `rustup`/`ghc`/`R` toolchains, files in `/tmp`, `/opt` edits | yes | **no** | **no** |
+
+Practical consequences:
+
+- **Durable work goes in `/workspace`** (it's the host filesystem, UID-aligned —
+  what you write appears with the user's normal ownership on the host).
+- **Runtime-installed system packages and language toolchains are ephemeral.**
+  If a task needs them reproducibly, it belongs in the image (Dockerfile) or a
+  project manifest, not an ad-hoc `apt install`. Tell the user when you install
+  something that won't survive.
+- **`~/.config/opencode` is a named volume** (`devbox-opencode-config`), so
+  things baked into the *image* under `/home/<user>/...` are **shadowed** by the
+  volume on existing containers and only seen on a fresh volume. This is exactly
+  why image-owned content that must always be live (the OMOS bundled skills, the
+  image-baked fallback skills, the harness instructions) lives under an image
+  path like `/usr/local/share/opencode-devbox/...` and is **symlinked** in by
+  the entrypoint on every start — never copied into a home dir a volume covers.
+  Copying image content into `~/.config/opencode` *freezes* it: a later
+  `docker compose pull` will not refresh it.
+
+## 2. Interactive shell vs. your tool shell (a real footgun)
+
+The conveniences below are defined in `~/.bash_aliases` and **only exist in an
+interactive login shell.** Your `bash` *tool* runs non-interactively, so these
+are "command not found" there — you must spell out the underlying command.
+
+| Interactive alias | Non-interactive equivalent to actually run |
+|---|---|
+| `dssh <host>` | `ssh -F "$HOME/.ssh-local/config" <host>` |
+| `dscp …` | `scp -F "$HOME/.ssh-local/config" …` |
+| `cat file` (→ `bat`) | `cat file` works, but output differs; use `command cat` for raw |
+| `ll`, `la` (→ `eza`/`ls`) | `ls -lh`, `ls -lha` |
+
+If a command "works in my terminal but not when the agent runs it," this alias
+gap is the first thing to suspect.
+
+**`dscp`/`scp` with accented filenames on a macOS host.** macOS stores filenames
+in Unicode **NFD** (decomposed — e.g. `ä` is `a` + combining U+0308), while the
+string you type or paste is usually **NFC** (precomposed `ä`, U+00E4). The bytes
+differ, so a precomposed remote path *silently* fails to match on the host —
+`scp … "mac:'~/Desktop/Skärmavbild ….png'"` returns *No such file or directory*
+even though the file plainly exists. Sidestep the encoding entirely: let the
+**remote shell expand a wildcard**, or list the directory first and copy the
+exact name it prints.
+
+```sh
+# glob dodges the NFC/NFD mismatch (the remote shell matches the real bytes):
+scp -F "$HOME/.ssh-local/config" "mac:~/Desktop/Sk*rmavbild*.png" ./
+# or read the exact filename first, then copy that:
+ssh -F "$HOME/.ssh-local/config" mac 'ls -1 ~/Desktop/*.png'
+```
+
+## 3. Reaching the Docker host and its LAN over SSH
+
+When the host is VM-backed (e.g. OrbStack / Docker Desktop on macOS) the
+entrypoint's `setup-lan-access.sh` writes a **writable SSH sidecar** at
+`~/.ssh-local/config`. It always provides:
+
+- A `Host *` block redirecting `ControlPath` into the writable `~/.ssh-local/cm`
+  (because `~/.ssh` is typically bind-mounted **read-only**, so a master socket
+  can't be created under it), plus `Include ~/.ssh/config`.
+- Aliases **`host` / `mac`** → `host.docker.internal` (user comes from
+  `HOST_SSH_USER`) — i.e. SSH back into the Docker host.
+- On VM-backed hosts only: an **SSH-jump-via-host** block so the container can
+  reach the host's directly-attached LAN peers (`ProxyJump host`). On a native
+  Linux host the LAN is usually reachable directly and this jump block is
+  omitted — **so don't assume a jump path exists; read the sidecar.**
+
+Use it (remember §2 — spell it out in tool bash):
+
+```sh
+ssh -F "$HOME/.ssh-local/config" mac 'hostname; whoami'      # reach the host
+ssh -F "$HOME/.ssh-local/config" <lan-peer> '…'             # reach a LAN peer (if configured)
+```
+
+Related mechanisms (don't reinvent them):
+
+- **ControlMaster multiplexing** is preconfigured (sockets under
+  `~/.ssh-local/cm`) to survive CGNAT per-destination flow caps on residential
+  ISPs. If `~/.ssh/config` pins a `ControlPath` under the read-only `~/.ssh`,
+  override with `-o ControlPath=none` (or use the sidecar, which already
+  redirects it).
+- To name a LAN peer (give it a stable alias + `ProxyJump host`), put the block
+  in the host-owned `~/.config/devbox-shell/ssh-lan.conf` (bind-mounted in and
+  `Include`d), **not** in the read-only `~/.ssh/config`. opencode has no
+  built-in "run my tools on a remote host" rewiring — the sidecar + plain `ssh`
+  is the path.
+
+## 4. DNS / name resolution — environment-specific, verify live
+
+How a name resolves here is **not universal** and depends on the host's
+networking. The container's own resolver is just `/etc/resolv.conf`, but the
+*host* (which you reach via §3, and whose DNS the container may inherit) can use
+**split-horizon DNS** to send certain internal domains to specific nameservers
+while everything else goes to a default resolver/VPN gateway. The mechanism is
+OS-specific and **may not be present at all**:
+
+- **macOS host:** per-domain files in `/etc/resolver/<domain>`, each listing
+  `nameserver` lines. Reading them (over `ssh … mac`) is a fine way to learn the
+  real split-DNS map — *for that one machine.*
+- **Linux host:** typically `systemd-resolved` split DNS (per-link `Domains=`
+  routing) or `/etc/resolv.conf` `search`/`nameserver`.
+- **Windows host:** the NRPT (Name Resolution Policy Table) plays the per-suffix
+  role; WSL2 inherits host resolution via mirrored networking + DNS tunneling.
+
+Operating rules:
+
+1. **Never hardcode a domain→nameserver mapping or a specific nameserver IP** —
+   it is per-deployment and changes between users and even VPN states.
+2. **Verify by reading the live config**, e.g. `cat /etc/resolv.conf` in the
+   container, or `ssh … mac 'cat /etc/resolver/* 2>/dev/null'` on a macOS host.
+3. **Reachability needs both DNS *and* a route.** A name resolving to an
+   internal address is useless if packets to that subnet don't have a path
+   (e.g. via the VPN or the §3 jump). Check both when something "resolves but
+   won't connect."
+4. If you discover deployment-specific facts (a domain, a nameserver, a
+   reachable peer), prefer recording them in MemPalace over baking them into
+   code or this skill.
+
+## 5. Python and other languages: uv-first, toolchains are ephemeral
+
+- A system `python3` exists, but **prefer `uv`** for REPLs and project envs —
+  it's installed and its store (`~/.local/share/uv`) is a persisted volume.
+  - Throwaway REPL: `uv run --with ipython ipython`
+  - Project env: `cd /workspace/proj && uv init && uv add <pkgs> && uv run …`
+    (the `pyproject.toml` + `uv.lock` travel with the repo — the durable choice).
+- Other language toolchains (Rust via rustup, R, GHC, Clojure, Go) are
+  **runtime opt-ins on the ephemeral layer** unless baked into the image — they
+  do not survive `down -v` or an image update. Flag this when installing.
+
+## 6. The oh-my-opencode-slim (OMOS) variant
+
+Present only in the `-omos` image (detection: `/usr/lib/node_modules/oh-my-opencode-slim`
+exists, or `ENABLE_OMOS=true`). It adds the bun runtime and multi-agent
+orchestration. Two environment facts matter:
+
+- **OMOS skills are symlinked from the image** into `~/.agents/skills/` by the
+  entrypoint on every start — they are **not** installed into the
+  `~/.config/opencode` volume (that froze them historically; see §1). A
+  `docker compose pull` + recreate refreshes them for free. Don't run the OMOS
+  installer with `--skills` to "fix" missing skills; the symlinks are the
+  mechanism.
+- Optional tmux orchestration is gated by `OMOS_TMUX`; the OMOS config lives at
+  `~/.config/opencode/oh-my-opencode-slim.json` and is regenerated only via
+  `OMOS_RESET=true`.
+
+## 7. MemPalace is the shared brain
+
+MemPalace data is usually a **host bind-mount** (or an optional `devbox-palace`
+named volume), so an opencode on the host and one in this container can share
+one palace (SQLite WAL: many readers, one writer). Use it to persist the
+deployment-specific facts this skill deliberately refuses to hardcode. Details
+are in the `mempalace` skill — and the harness instruction shipped with this
+image points you at it at session start.
+
+## Checklist before acting in this environment
+
+- [ ] Writing durable output? → `/workspace`, not the ephemeral layer.
+- [ ] Using `dssh`/`dscp`/`ll` in the bash tool? → spell out the real command.
+- [ ] Assuming a hostname / domain / nameserver / host OS? → stop, detect it.
+- [ ] "Resolves but won't connect"? → check route *and* DNS (§3 + §4).
+- [ ] `apt`/toolchain install? → tell the user it's ephemeral unless imaged.
+- [ ] Shipping image-owned content opencode reads from `~/.config/opencode`? →
+      symlink it from `/usr/...`, never copy into the volume (§1).

rootfs/usr/local/share/terminfo-src/ghostty.terminfo

@@ -0,0 +1,14 @@
+# xterm-ghostty — alias of the maintained ncurses `ghostty` terminfo entry.
+#
+# Ghostty sets TERM=xterm-ghostty by default, but the ncurses terminfo
+# database (Debian: ncurses-term) ships the entry under the name `ghostty`
+# only — there is no `xterm-ghostty` alias, and no distro packages one. This
+# thin alias makes xterm-ghostty resolve to the same upstream-maintained
+# capability set, so SSH sessions from a Ghostty terminal work without
+# vendoring Ghostty's full (Zig-generated) terminfo here.
+#
+# `use=ghostty` is resolved by `tic` at compile time against the base
+# `ghostty` entry from ncurses-term (installed in Dockerfile.base before the
+# compile step). Compiled with `tic -x`.
+xterm-ghostty|Ghostty terminal emulator (xterm-ghostty alias),
+	use=ghostty,

scripts/check-base-hash.sh

@@ -0,0 +1,43 @@
+#!/usr/bin/env bash
+# check-base-hash.sh — guard the base-rebuild invariant.
+#
+# Every floating `ARG *_REF` consumed by Dockerfile.base MUST be folded
+# into the base_tag hash in the docker-publish workflow. Otherwise a
+# ref-only change to that dependency does not change the base hash, the
+# Docker Hub probe finds the old base tag, and the base is NOT rebuilt —
+# the dependency fix silently fails to land. This is the v1.1.2-class
+# staleness footgun (then it was mempalace-toolkit; this guard stops the
+# next one before it ships).
+#
+# Runs in CI (base-decide job) and locally:  bash scripts/check-base-hash.sh
+set -euo pipefail
+cd "$(dirname "$0")/.."
+
+WF=".gitea/workflows/docker-publish-split.yml"
+DF="Dockerfile.base"
+
+# Extract the hash-compute block: the `HASH=$( … ) | sha256sum | cut`
+# brace-group in the "Compute base tag" step. This lives in a separate
+# file from the workflow, so scanning $WF here is free of the self-match
+# hazard an inline workflow step would have.
+block=$(awk '/HASH=\$\(/{f=1} f{print} f && /cut -c1-12/{exit}' "$WF")
+if [ -z "$block" ]; then
+  echo "::error::could not locate the HASH=\$( … ) | sha256sum block in $WF"
+  exit 1
+fi
+
+refs=$(grep -oE '^ARG [A-Z0-9_]+_REF' "$DF" | awk '{print $2}' | sort -u)
+fail=0
+for r in $refs; do
+  lc=$(printf '%s' "$r" | tr '[:upper:]' '[:lower:]')
+  if ! printf '%s' "$block" | grep -q "outputs.$lc"; then
+    echo "::error::Dockerfile.base declares '$r' but it is NOT folded into the base_tag hash in $WF."
+    echo "::error::Add  echo \"\${{ needs.resolve-versions.outputs.$lc }}\"  inside the HASH=\$( … ) | sha256sum block, or a $r-only change will silently fail to rebuild the base."
+    fail=1
+  fi
+done
+
+if [ "$fail" = 0 ]; then
+  echo "OK: all Dockerfile.base *_REF args are folded into base_tag (${refs:-none})."
+fi
+exit $fail

scripts/check-workflow-shell.sh

@@ -0,0 +1,67 @@
+#!/usr/bin/env bash
+# Gitea-accurate guard against the "bash syntax under the default sh/dash
+# shell" footgun. Ported from pi-devbox, where this class bit twice
+# (ed49b8d resolve-versions; b7197e8/b33e9dc promote-base-latest, run 418).
+# opencode-devbox has not been bitten yet — this is a PREVENTIVE guard so a
+# future author can't reintroduce the class.
+#
+# WHY A CUSTOM CHECK AND NOT JUST actionlint:
+#   actionlint models *GitHub* Actions, whose default `run` shell is bash. It
+#   therefore assumes a step that omits `shell:` runs under bash, and does NOT
+#   flag `set -o pipefail` there. Gitea Actions' default is `sh` (dash), so the
+#   exact bug (omit `shell:`, use bash syntax) is invisible to actionlint.
+#   actionlint only fires when a step *explicitly* declares `shell: sh`.
+#
+# THE INVARIANT THIS ENFORCES:
+#   Every `run:` step in every .gitea/workflows/*.yml must resolve to an
+#   effective shell of `bash` — via the step's own `shell:`, a job-level
+#   `defaults.run.shell`, or a workflow-level `defaults.run.shell`. Any step
+#   that would fall through to Gitea's `sh` default is a FAILURE, because a
+#   future author adding bash syntax to it fails silently in CI.
+#
+# Pair this with actionlint (which catches explicit `shell: sh` + bash syntax,
+# expression errors, and much else). Together they cover the class on Gitea.
+set -euo pipefail
+
+WF_DIR="${1:-.gitea/workflows}"
+
+python3 - "$WF_DIR" <<'PY'
+import sys, glob, os
+try:
+    import yaml
+except ImportError:
+    sys.stderr.write("ERROR: python3 yaml module missing (apt install python3-yaml)\n")
+    sys.exit(2)
+
+wf_dir = sys.argv[1]
+files = sorted(glob.glob(os.path.join(wf_dir, "*.yml")) + glob.glob(os.path.join(wf_dir, "*.yaml")))
+if not files:
+    sys.stderr.write(f"ERROR: no workflow files under {wf_dir}\n")
+    sys.exit(2)
+
+problems = []
+for f in files:
+    with open(f) as fh:
+        doc = yaml.safe_load(fh) or {}
+    wf_shell = (((doc.get("defaults") or {}).get("run") or {}).get("shell"))
+    jobs = doc.get("jobs") or {}
+    for jname, job in jobs.items():
+        job = job or {}
+        job_shell = (((job.get("defaults") or {}).get("run") or {}).get("shell"))
+        steps = job.get("steps") or []
+        for i, step in enumerate(steps):
+            step = step or {}
+            if "run" not in step:
+                continue  # `uses:` steps have no shell
+            eff = step.get("shell") or job_shell or wf_shell or "sh"  # Gitea default = sh
+            if eff != "bash":
+                name = step.get("name") or f"step[{i}]"
+                problems.append(f"{f}: job '{jname}' / '{name}': effective shell = '{eff}' (Gitea default is sh; declare shell: bash or a bash default)")
+
+if problems:
+    sys.stderr.write("Workflow shell guard FAILED — bash default not guaranteed:\n")
+    for p in problems:
+        sys.stderr.write(f"  - {p}\n")
+    sys.exit(1)
+print(f"Workflow shell guard OK — all run: steps in {len(files)} workflow file(s) resolve to bash.")
+PY

scripts/generate-dockerhub-md.py

@@ -56,23 +56,24 @@ HUB_TEMPLATE = f"""# opencode-devbox
 
 Portable AI developer environment for [opencode](https://opencode.ai). Debian-based, with git, SSH, Node.js, AWS CLI v2, and common dev tools pre-installed.
 
+> **Current `:latest` ships opencode `{{{{OPENCODE_VERSION}}}}`** (the baked version is asserted by smoke tests, so this page never drifts from the image).
+
 Designed for teams who want a reproducible coding-agent setup that runs the same on every laptop and CI runner — without forcing each developer to install Bun, Node, AWS CLI, mempalace, or maintain shell config drift across machines.
 
 ## Image Variants
 
 | Tag | Description |
 |---|---|
-| `latest` / `vX.Y.Z` | Base image — opencode, Node.js, AWS CLI, dev tools |
+| `latest` / `vX.Y.Z` | Base image — opencode `{{{{OPENCODE_VERSION}}}}`, Node.js, AWS CLI, dev tools |
 | `latest-omos` / `vX.Y.Z-omos` | Base + [oh-my-opencode-slim](https://github.com/alvinunreal/oh-my-opencode-slim) multi-agent orchestration and Bun |
-| `latest-with-pi` / `vX.Y.Z-with-pi` | Base + [pi](https://github.com/earendil-works/pi) as alternative/complementary harness (shares the mempalace install with opencode) |
-| `latest-omos-with-pi` / `vX.Y.Z-omos-with-pi` | OMOS + pi together |
 
 All variants support `linux/amd64` and `linux/arm64`.
 
-> A fifth, pi-without-opencode build is produced from the same `Dockerfile.variant`
-> (`INSTALL_OPENCODE=false`) but is **not** published under this repo — it ships as
-> the separate [`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox)
-> image so an "opencode-devbox" tag never lacks opencode.
+> **Looking for pi?** As of v2.0.0 the pi coding-agent is no longer bundled in
+> opencode-devbox. It ships as the dedicated
+> [`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox) image, which
+> shares the same mempalace memory layer. See
+> <https://gitea.jordbo.se/joakimp/pi-devbox>.
 
 ## Quick Start
 
@@ -87,7 +88,7 @@ curl -fsSL https://gitea.jordbo.se/joakimp/opencode-devbox/raw/branch/main/.env.
 docker compose run --rm devbox
 ```
 
-This drops you straight into opencode with your project mounted at `/workspace`. Use `bash` as the command (e.g. `docker compose run --rm devbox bash`) to land in a shell first — useful for `aws sso login`, `pi` (on `*-with-pi` variants), or multi-harness workflows.
+This mounts your project at `/workspace`. With no command (as above) the image's default `CMD` (`bash -l`) drops you into a login shell — run `opencode` to start the harness, or `aws sso login` first, etc. To start opencode directly, pass it as the command: `docker compose run --rm devbox opencode`.
 
 **One-shot run, no persistence:**
 
@@ -107,11 +108,10 @@ Full setup guide — authentication for each provider (Anthropic, OpenAI, Bedroc
 ## What's Inside
 
 - **[opencode](https://opencode.ai)** — primary coding-agent harness. Multi-provider (Anthropic, OpenAI, Bedrock, Google, Groq, etc.).
-- **[pi](https://github.com/earendil-works/pi)** *(in `*-with-pi` variants)* — lightweight TUI coding-agent that coexists with opencode and shares the same mempalace install. Includes the `mcp-loader` extension so any local-stdio or remote streamable-HTTP MCP server (searxng, gitea, context7, …) can be added by editing `~/.pi/agent/settings.json`.
-- **[mempalace](https://github.com/MemPalace/mempalace)** — persistent AI memory layer (ChromaDB + SQLite). Wing/diary/knowledge-graph entries are mutually visible to opencode and pi.
+- **[mempalace](https://github.com/MemPalace/mempalace)** — persistent AI memory layer (ChromaDB + SQLite). Wing/diary/knowledge-graph entries are shareable with the sibling [`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox) image when both point at the same palace.
 - **[oh-my-opencode-slim](https://github.com/alvinunreal/oh-my-opencode-slim)** *(in `*-omos` variants)* — multi-agent orchestration on top of opencode (council, fallback chains, named agents).
 - **AWS CLI v2** with SSO support, **Node.js LTS**, **Bun** (OMOS variants), **uv** (Python), **gosu** for clean UID/GID adjustment to match your host workspace.
-- **MCP wrappers** for mempalace pre-installed and pre-wired to both harnesses.
+- **MCP wrappers** for mempalace pre-installed and pre-wired to opencode.
 
 ## Authentication
 
@@ -129,8 +129,7 @@ Full Bedrock walkthrough (IAM roles, permissions, multi-account setups): see the
 
 | Volume | Mount | Survives |
 |---|---|---|
-| `devbox-opencode-config` | `~/.config/opencode` | container recreate, image rebuild |
-| `devbox-pi-config` | `~/.pi` | container recreate, image rebuild — incl. user-installed pi packages via `pi install` (`NPM_CONFIG_PREFIX` points into the volume) |
+| `devbox-opencode-config` | `~/.config/opencode` | container recreate, image rebuild — incl. user-installed npm globals via `npm install -g` (`NPM_CONFIG_PREFIX` points into the volume) |
 | `devbox-palace` (uncomment) | `~/.mempalace` | container recreate, image rebuild — palace data is precious, treat as primary storage |
 | `devbox-chroma-cache` | `~/.cache/chroma` | container recreate (model cache, disposable — re-downloads in seconds) |
 
@@ -147,7 +146,7 @@ Full persistence reference, including multi-user (`SIGNUM`) isolation and host b
 
 ## Sibling images
 
-- **[`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox)** — pi-only image built on top of this image's base layer. Smaller (~700 MB) and version-tracks the [pi npm package](https://www.npmjs.com/package/@earendil-works/pi-coding-agent) directly. Use this if you want pi without opencode. Source: <https://gitea.jordbo.se/joakimp/pi-devbox>
+- **[`joakimp/pi-devbox`](https://hub.docker.com/r/joakimp/pi-devbox)** — the pi coding-agent in its own self-contained image, built on a shared Debian base. Version-tracks the [pi npm package](https://www.npmjs.com/package/@earendil-works/pi-coding-agent) directly and can share this image's mempalace palace. Use it if you want pi instead of (or alongside) opencode. Source: <https://gitea.jordbo.se/joakimp/pi-devbox>
 
 ## License
 
@@ -155,7 +154,7 @@ MIT. See <{GITEA}/src/branch/main/LICENSE>.
 
 ---
 
-> This description is generated by `scripts/generate-dockerhub-md.py` from a hand-maintained template. Edit the template (not this file) and regenerate.
+> This description is generated by `scripts/generate-dockerhub-md.py` from a hand-maintained template. Edit the template (not this file) and regenerate. The `{{{{OPENCODE_VERSION}}}}` placeholder is filled by CI at publish time.
 """
 
 

scripts/recreate-sanity-check.sh

@@ -0,0 +1,234 @@
+#!/usr/bin/env bash
+# Runtime post-recreate verification for opencode-devbox.
+#
+# Verifies that after `docker compose up -d --force-recreate`:
+#   - The new image is actually live (opencode version matches Dockerfile.variant)
+#   - Persisted named volumes survived (mempalace palace, opencode.db, bash-history)
+#   - OMOS runtime skill symlinks resolve (omos variant only)
+#   - Shell defaults re-seeded from /etc/skel-devbox
+#   - /opt toolkits intact
+#   - Known expected-absences don't regress
+#
+# This is repo/maintainer tooling — the runtime peer of smoke-test.sh. It is
+# NOT baked into the published Docker Hub image; run it from a checkout of the
+# opencode-devbox repo (which a maintainer already has for CI builds). A plain
+# `docker pull` consumer is not the audience and will not have this file.
+#
+# Usage: ./scripts/recreate-sanity-check.sh [--expected-version X.Y.Z] [--variant base|omos]
+#
+# Exit codes:
+#   0  all checks passed
+#   1  one or more checks failed
+#   2  usage error
+
+set -euo pipefail
+
+EXPECTED_VERSION=""
+VARIANT=""
+REPO_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
+
+# Parse arguments
+while [[ $# -gt 0 ]]; do
+  case "$1" in
+    --expected-version)
+      EXPECTED_VERSION="$2"
+      shift 2
+      ;;
+    --variant)
+      VARIANT="$2"
+      shift 2
+      ;;
+    *)
+      echo "usage: $0 [--expected-version X.Y.Z] [--variant base|omos]" >&2
+      exit 2
+      ;;
+  esac
+done
+
+FAILED=0
+pass() { echo "  ✓ $1"; }
+fail() { echo "  ✗ $1" >&2; FAILED=$((FAILED + 1)); }
+warn() { echo "  ⚠ $1" >&2; }
+
+# Determine expected opencode version from Dockerfile.variant if not provided
+if [ -z "$EXPECTED_VERSION" ]; then
+  EXPECTED_VERSION="$(grep -oE 'OPENCODE_VERSION=[0-9.]+' "$REPO_DIR/Dockerfile.variant" | head -1 | cut -d= -f2)"
+  if [ -z "$EXPECTED_VERSION" ]; then
+    echo "error: could not determine OPENCODE_VERSION from $REPO_DIR/Dockerfile.variant" >&2
+    exit 2
+  fi
+fi
+
+# Auto-detect variant if not provided
+if [ -z "$VARIANT" ]; then
+  if command -v bun >/dev/null 2>&1 || [ -d /usr/lib/node_modules/oh-my-opencode-slim ] || [ -d /usr/local/lib/node_modules/oh-my-opencode-slim ]; then
+    VARIANT="omos"
+  else
+    VARIANT="base"
+  fi
+fi
+
+# Print header with git context
+echo "=== Recreate sanity check (variant: $VARIANT) ==="
+if GIT_TAG=$(git -C "$REPO_DIR" describe --tags 2>/dev/null); then
+  echo "  Repo HEAD: $GIT_TAG (version-match only meaningful when image tag matches)"
+else
+  echo "  Repo HEAD: (not a git repo or no tags)"
+fi
+echo
+
+echo "-- opencode version --"
+if ACTUAL_VERSION=$(opencode --version 2>&1 | head -1); then
+  if [ "$ACTUAL_VERSION" = "$EXPECTED_VERSION" ]; then
+    pass "opencode version $ACTUAL_VERSION"
+  else
+    fail "opencode version mismatch: expected $EXPECTED_VERSION, got $ACTUAL_VERSION"
+  fi
+else
+  fail "opencode --version failed"
+fi
+
+echo
+echo "-- Persisted named volumes (must survive --force-recreate) --"
+
+# mempalace palace volume
+if [ -f "$HOME/.mempalace/palace/chroma.sqlite3" ]; then
+  SIZE=$(du -h "$HOME/.mempalace/palace/chroma.sqlite3" | cut -f1)
+  if [ -s "$HOME/.mempalace/palace/chroma.sqlite3" ]; then
+    pass "~/.mempalace/palace/chroma.sqlite3 exists ($SIZE)"
+  else
+    fail "~/.mempalace/palace/chroma.sqlite3 exists but is empty"
+  fi
+else
+  fail "~/.mempalace/palace/chroma.sqlite3 missing"
+fi
+
+# opencode session history volume
+if [ -f "$HOME/.local/share/opencode/opencode.db" ]; then
+  SIZE=$(du -h "$HOME/.local/share/opencode/opencode.db" | cut -f1)
+  if [ -s "$HOME/.local/share/opencode/opencode.db" ]; then
+    pass "~/.local/share/opencode/opencode.db exists ($SIZE)"
+  else
+    fail "~/.local/share/opencode/opencode.db exists but is empty"
+  fi
+else
+  fail "~/.local/share/opencode/opencode.db missing"
+fi
+
+# bash-history volume mount point (empty .bash_history right after recreate is NORMAL)
+if [ -d "$HOME/.cache/bash" ]; then
+  pass "~/.cache/bash exists as directory"
+else
+  fail "~/.cache/bash missing or not a directory"
+fi
+
+echo
+echo "-- omos runtime skill symlinks (omos variant only; skip on base) --"
+if [ "$VARIANT" = "omos" ]; then
+  SKILLS_OK=0
+  SKILLS_TOTAL=5
+  for skill in clonedeps codemap deepwork oh-my-opencode-slim simplify; do
+    SKILL_PATH="$HOME/.agents/skills/$skill"
+    if [ -L "$SKILL_PATH" ]; then
+      TARGET=$(readlink -f "$SKILL_PATH")
+      # Check if target resolves to a real directory and contains the expected path
+      if [ -d "$TARGET" ] && echo "$TARGET" | grep -q "node_modules/oh-my-opencode-slim/src/skills/$skill"; then
+        SKILLS_OK=$((SKILLS_OK + 1))
+      else
+        fail "~/.agents/skills/$skill symlink target invalid: $TARGET"
+      fi
+    else
+      fail "~/.agents/skills/$skill missing or not a symlink"
+    fi
+  done
+  if [ "$SKILLS_OK" -eq "$SKILLS_TOTAL" ]; then
+    pass "$SKILLS_OK/$SKILLS_TOTAL omos skill symlinks resolve"
+  fi
+
+  # Migration marker
+  if [ -f "$HOME/.config/opencode/.omos-skills-migrated" ]; then
+    pass "~/.config/opencode/.omos-skills-migrated exists"
+  else
+    fail "~/.config/opencode/.omos-skills-migrated missing"
+  fi
+else
+  echo "  - skipped (base variant)"
+fi
+
+echo
+echo "-- Image-baked fallback skills + harness instruction (both variants) --"
+# Baked under /usr/local/share and linked in by entrypoint-user.sh.
+# opencode-devbox-environment uses a name unlikely to be overridden, so it is a
+# reliable probe that the skills reconcile ran; the instruction confirms the
+# ~/.config/opencode/instructions symlink survived the named volume.
+if [ -e "$HOME/.agents/skills/opencode-devbox-environment/SKILL.md" ]; then
+  pass "~/.agents/skills/opencode-devbox-environment resolves"
+else
+  fail "~/.agents/skills/opencode-devbox-environment missing"
+fi
+if [ -e "$HOME/.agents/skills/mempalace/SKILL.md" ]; then
+  pass "~/.agents/skills/mempalace resolves"
+else
+  fail "~/.agents/skills/mempalace missing"
+fi
+if [ -e "$HOME/.config/opencode/instructions/opencode-devbox.md" ]; then
+  pass "~/.config/opencode/instructions/opencode-devbox.md resolves"
+else
+  fail "~/.config/opencode/instructions/opencode-devbox.md missing"
+fi
+
+echo
+echo "-- Shell defaults re-seeded from /etc/skel-devbox --"
+if [ -f "$HOME/.bash_aliases" ]; then
+  pass "~/.bash_aliases exists"
+else
+  fail "~/.bash_aliases missing"
+fi
+
+if [ -f "$HOME/.inputrc" ]; then
+  pass "~/.inputrc exists"
+else
+  fail "~/.inputrc missing"
+fi
+
+echo
+echo "-- cli_utils bind-mount --"
+if [ -d /workspace/cli_utils ] && [ -d /workspace/cli_utils/.git ]; then
+  pass "/workspace/cli_utils exists with .git subdir"
+else
+  fail "/workspace/cli_utils missing or .git subdir absent"
+fi
+
+echo
+echo "-- Baked /opt toolkits --"
+if [ -d /opt/mempalace-toolkit ]; then
+  if MEMPALACE_SESSION_PATH=$(command -v mempalace-session 2>/dev/null); then
+    RESOLVED=$(readlink -f "$MEMPALACE_SESSION_PATH")
+    pass "/opt/mempalace-toolkit exists, mempalace-session resolves to $RESOLVED"
+  else
+    fail "/opt/mempalace-toolkit exists but mempalace-session not on PATH"
+  fi
+else
+  fail "/opt/mempalace-toolkit missing"
+fi
+
+echo
+echo "-- Known expected-absences (regressions vs by-design) --"
+if [ ! -d "$HOME/.local/bin" ]; then
+  warn "~/.local/bin absent — expected; mempalace toolkit relocated to /opt (not a wrapper-loss regression)"
+else
+  pass "~/.local/bin exists (toolkit may have been installed locally)"
+fi
+
+if ! command -v go >/dev/null 2>&1; then
+  warn "go absent — expected unless image built with INSTALL_GO=true"
+else
+  pass "go is on PATH"
+fi
+
+echo
+if [ "$FAILED" -gt 0 ]; then
+  echo "=== FAILED: $FAILED check(s) ===" >&2
+  exit 1
+fi
+echo "=== PASSED ==="

scripts/smoke-test.sh

@@ -3,12 +3,17 @@
 #
 # Verifies:
 #   - Core binaries are on PATH and runnable
+#   - non-modal editors nano + micro are present (alongside nvim)
 #   - opencode itself starts and prints a version
 #   - Entrypoint runs cleanly as non-root after UID adjustment
 #   - Generated opencode.json has the expected shape
 #   - MCP wrapper works (when mempalace is installed)
+#   - typst PDF engine for pandoc — `pandoc --pdf-engine=typst`
+#   - terminfo for modern emulators: xterm-kitty, xterm-ghostty, wezterm,
+#     alacritty, foot (kitty-terminfo + ncurses-term + compiled ghostty alias)
+#   - Neovim true-colour default (sysinit.vim termguicolors)
 #
-# Usage: ./scripts/smoke-test.sh <image> [--variant base|omos|with-pi|omos-with-pi|pi-only]
+# Usage: ./scripts/smoke-test.sh <image> [--variant base|omos]
 #
 # Exit codes:
 #   0  all checks passed
@@ -23,13 +28,20 @@ if [ "${2:-}" = "--variant" ]; then
 fi
 
 if [ -z "$IMAGE" ]; then
-  echo "usage: $0 <image> [--variant base|omos|with-pi|omos-with-pi|pi-only]" >&2
+  echo "usage: $0 <image> [--variant base|omos]" >&2
   exit 2
 fi
 
 FAILED=0
 pass() { echo "  ✓ $1"; }
 fail() { echo "  ✗ $1" >&2; FAILED=$((FAILED + 1)); }
+warn() { echo "  ⚠ $1" >&2; }
+
+# Registration assertions for fork/recall were removed in v2.0.0 along with
+# pi. STRICT_REGISTRATION is retained as an inert env var for backward
+# compatibility with any external caller that still sets it; it has no
+# effect now that no pi packages are deployed.
+STRICT_REGISTRATION="${STRICT_REGISTRATION:-0}"
 
 run() {
   # Run a command inside the image and capture its output.
@@ -72,9 +84,6 @@ docker run --rm --entrypoint="" "$IMAGE" sh -c '
   if command -v opencode >/dev/null 2>&1; then
     printf "  %-15s %s\n" "opencode"      "$(opencode --version 2>&1 | head -1)"
   fi
-  if command -v pi >/dev/null 2>&1; then
-    printf "  %-15s %s\n" "pi"            "$(pi --version 2>&1 | head -1)"
-  fi
   printf "  %-15s %s\n" "node"          "$(node --version)"
   printf "  %-15s %s\n" "npm"           "$(npm --version)"
   printf "  %-15s %s\n" "nvim"          "$(nvim --version | head -1)"
@@ -104,7 +113,7 @@ docker run --rm --entrypoint="" "$IMAGE" sh -c '
 echo
 echo "-- Core binaries --"
 # opencode is gated on INSTALL_OPENCODE=true (default). When absent, the
-# image is a pi-only build (or a pure base — no harness at all).
+# image is a pure base with no harness at all.
 if docker run --rm --entrypoint="" "$IMAGE" sh -c "command -v opencode" >/dev/null 2>&1; then
   run "opencode"          "opencode --version"
 else
@@ -114,6 +123,12 @@ run "node"              "node --version"
 run "npm"               "npm --version"
 run "git"               "git --version"
 run "nvim"              "nvim --version | head -1"
+run "nvim true-colour default (sysinit.vim)" "nvim --headless -c 'lua os.exit(vim.o.termguicolors and 0 or 1)'"
+run "kitty-terminfo"    "infocmp -x xterm-kitty >/dev/null 2>&1"
+run "terminfo: modern emulators (ncurses-term)" 'for t in wezterm alacritty foot ghostty st-256color; do infocmp -x "$t" >/dev/null 2>&1 || exit 1; done'
+run "terminfo: xterm-ghostty alias (tic)" "infocmp -x xterm-ghostty >/dev/null 2>&1"
+run "nano"              "nano --version | head -1"
+run "micro"             "micro --version"
 run "bat"               "bat --version"
 run "eza"               "eza --version | head -1"
 run "zoxide"            "zoxide --version"
@@ -124,12 +139,44 @@ run "fzf"               "fzf --version"
 run "fd"                "fd --version"
 run "rg"                "rg --version | head -1"
 run "jq"                "jq --version"
+# yq MUST be mikefarah's Go yq v4, NOT Debian's `yq` apt package (the unrelated
+# Python kislyuk/yq — a jq wrapper on a 3.x line with incompatible syntax). v2.9.0
+# swapped the apt package for the mikefarah binary. Pinning the contract to major
+# v4 makes BOTH a regression to the Python package AND a surprise future yq v5
+# fail CI loudly, instead of silently breaking the cloud-init repo's
+# provision.sh/deploy.sh which require v4 syntax.
+run "yq is mikefarah v4"  "yq --version | grep -qE 'mikefarah.*version v4' && yq --version"
 run "git-crypt"         "git-crypt --version | head -1"
 run "gitleaks"          "gitleaks version"
 run "aws"               "aws --version"
 run "gitea-mcp"         "gitea-mcp --version"
 run "gosu"              "gosu --version"
 run "tmux"              "tmux -V"
+run "pandoc"            "pandoc --version | head -1"
+run "typst"             "typst --version"
+# agent-browser + its Chromium. The ENV must point at a resolvable executable:
+# AGENT_BROWSER_EXECUTABLE_PATH -> /usr/local/bin/agent-chrome -> Playwright's
+# per-version, per-ARCH chrome binary. Asserting the resolved target is
+# executable catches a Playwright layout change (the reason the Dockerfile
+# `find`s the binary instead of hardcoding the path) rather than just checking
+# that a dangling symlink exists.
+run "agent-browser"      "agent-browser --version"
+run "agent-chrome resolves to an executable" \
+    "test -x \"\$(readlink -f /usr/local/bin/agent-chrome)\" && readlink -f /usr/local/bin/agent-chrome"
+run_expect "AGENT_BROWSER_EXECUTABLE_PATH baked" \
+    "printenv AGENT_BROWSER_EXECUTABLE_PATH" "/usr/local/bin/agent-chrome"
+run "pandoc+typst PDF engine" "printf '# hi\n' | pandoc --pdf-engine=typst -o /tmp/_smoke.pdf - && test -s /tmp/_smoke.pdf; rm -f /tmp/_smoke.pdf"
+run "graphviz (dot)"    "dot -V"
+run "tldr (tealdeer)"   "tldr --version"
+run "dot-watch"         "test -x /usr/local/bin/dot-watch && bash -n /usr/local/bin/dot-watch && echo ok"
+
+# Background subagents: opencode gates them behind this experimental env var,
+# and OMOS V2+ default orchestration depends on it. Baked ON as an ENV in
+# Dockerfile.base — assert it's present in the image environment (both variants).
+run_expect "bg-subagents env baked" "printenv OPENCODE_EXPERIMENTAL_BACKGROUND_SUBAGENTS" "true"
+# True-colour advertisement for colour-aware tools (bat, delta, Neovim's
+# auto-detect). Pairs with the terminfo entries + sysinit.vim termguicolors.
+run_expect "COLORTERM env baked" "printenv COLORTERM" "truecolor"
 
 # SSH ControlMaster baked defaults: the config file must exist (image-level)
 # and ssh -G must report ControlPath rooted at /tmp/sshcm/ for an arbitrary
@@ -158,92 +205,15 @@ elif docker run --rm --entrypoint="" "$IMAGE" sh -c "command -v mempalace" >/dev
   echo "  - mempalace-toolkit not installed (INSTALL_MEMPALACE_TOOLKIT=false)"
 fi
 
-# pi: present when built with INSTALL_PI=true. Verifies pi itself plus
-# the runtime-deployed pi-toolkit + pi-extensions + mempalace bridge
-# symlinks under ~/.pi/agent/. Note: extension symlinks are created by
-# entrypoint-user.sh on first start, so we test by running the entry
-# point chain (not just `docker run --entrypoint=""`).
-if docker run --rm --entrypoint="" "$IMAGE" sh -c "command -v pi" >/dev/null 2>&1; then
-  if [ -n "${EXPECTED_PI_VERSION:-}" ]; then
-    run_expect "pi version matches build-arg" "pi --version" "$EXPECTED_PI_VERSION"
-  else
-    run "pi"                       "pi --version"
-  fi
-  run "pi-toolkit clone"         "test -d /opt/pi-toolkit && git -C /opt/pi-toolkit rev-parse --short HEAD"
-  run "pi-extensions clone"      "test -d /opt/pi-extensions && git -C /opt/pi-extensions rev-parse --short HEAD"
-  # pi-fork (fork tool) + pi-observational-memory (recall tool): cloned to
-  # /opt with node_modules baked at build time (a local-path `pi install` does
-  # NOT npm-install, so deps MUST already be present for the extension to load).
-  run "pi-fork clone + node_modules" \
-      "test -f /opt/pi-fork/package.json && test -d /opt/pi-fork/node_modules && echo ok"
-  run "pi-observational-memory clone + node_modules" \
-      "test -f /opt/pi-observational-memory/package.json && test -d /opt/pi-observational-memory/node_modules && echo ok"
-
-  # Run the full entrypoint as developer to verify install.sh deployment.
-  # Spin up a long-running container so we can `docker exec` into it from
-  # the host — the `run` helper above invokes commands INSIDE the image
-  # and has no docker CLI to nest with.
-  CID=$(docker run -d --rm "$IMAGE" tail -f /dev/null)
-  trap 'docker rm -f "$CID" >/dev/null 2>&1 || true' EXIT
-
-  # Wait for entrypoint-user.sh to finish deploying pi-toolkit + extensions.
-  # Marker: keybindings.json symlink lands once pi-toolkit/install.sh has run.
-  # Up to 30s — omos-with-pi has more setup work than base+pi.
-  for _ in $(seq 1 30); do
-    if docker exec "$CID" test -L /home/developer/.pi/agent/keybindings.json 2>/dev/null; then
-      break
-    fi
-    sleep 1
-  done
-
-  exec_test() {
-    local label="$1"; shift
-    local out
-    if out=$(docker exec -u developer "$CID" sh -c "$*" 2>&1); then
-      pass "$label ($(echo "$out" | head -1))"
-    else
-      fail "$label: $out"
-    fi
-  }
-
-  exec_test "~/.pi/agent/keybindings.json (pi-toolkit)" \
-            'test -L $HOME/.pi/agent/keybindings.json && echo ok'
-  exec_test "~/.pi/agent/extensions/*.ts ≥ 4 (pi-extensions)" \
-            'count=$(ls -1 $HOME/.pi/agent/extensions/*.ts 2>/dev/null | wc -l); [ $count -ge 4 ] && echo "$count extensions"'
-  exec_test "~/.pi/agent/extensions/mempalace.ts (bridge)" \
-            'test -L $HOME/.pi/agent/extensions/mempalace.ts && echo ok'
-  exec_test "~/.pi/agent/settings.json (template bootstrap)" \
-            'test -f $HOME/.pi/agent/settings.json && echo ok'
-
-  # pi-fork + pi-observational-memory are registered by entrypoint-user.sh via
-  # `pi install /opt/<pkg>` (records a relative path into settings.json
-  # packages). That runs slightly after the keybindings marker, so wait for it.
-  for _ in $(seq 1 15); do
-    if docker exec "$CID" grep -q pi-observational-memory \
-         /home/developer/.pi/agent/settings.json 2>/dev/null; then
-      break
-    fi
-    sleep 1
-  done
-  exec_test "pi-fork registered in settings.json (fork tool)" \
-            'grep -q pi-fork $HOME/.pi/agent/settings.json && echo ok'
-  exec_test "pi-observational-memory registered in settings.json (recall tool)" \
-            'grep -q pi-observational-memory $HOME/.pi/agent/settings.json && echo ok'
-
-  docker rm -f "$CID" >/dev/null 2>&1 || true
-  trap - EXIT
-else
-  echo "  - pi not installed (INSTALL_PI=false)"
-fi
-
-# bun: only in the omos and omos-with-pi variants
-if [ "$VARIANT" = "omos" ] || [ "$VARIANT" = "omos-with-pi" ]; then
+# bun: only in the omos variant
+if [ "$VARIANT" = "omos" ]; then
   run "bun (omos)"            "bun --version"
   run "bunx symlink (omos)"   "test -L /usr/local/bin/bunx && readlink /usr/local/bin/bunx"
   # oh-my-opencode-slim is npm-installed globally (not a bun install);
   # verify it shows up in the global module list. We must explicitly point
   # npm at the system prefix (/usr) here: the image's NPM_CONFIG_PREFIX env
-  # is set to /home/developer/.pi/npm-global so user-installed packages
+  # is set to /home/developer/.config/opencode/npm-global so user-installed
+  # packages
   # land on the persistent volume — which means a default `npm ls -g`
   # queries the user prefix and would miss the baked binaries even though
   # they're correctly on PATH at /usr/bin.
@@ -253,6 +223,15 @@ if [ "$VARIANT" = "omos" ] || [ "$VARIANT" = "omos-with-pi" ]; then
                "NPM_CONFIG_PREFIX=/usr npm ls -g --depth=0 2>/dev/null | grep oh-my-opencode-slim" \
                "$EXPECTED_OMOS_VERSION"
   fi
+  # OMOS bundled-skills SOURCE must be present at the fixed image path that
+  # entrypoint-user.sh symlinks into ~/.agents/skills/ on container start. If
+  # upstream restructures the package (moves src/skills), the runtime symlinks
+  # would dangle SILENTLY and the skills would just disappear — assert the
+  # source here so that breakage fails the build loudly instead. We check the
+  # source (not the runtime symlinks) because smoke tests run with
+  # --entrypoint="" and never execute entrypoint-user.sh.
+  run "omos bundled-skills source" \
+      "for n in clonedeps codemap deepwork oh-my-opencode-slim simplify; do test -d /usr/lib/node_modules/oh-my-opencode-slim/src/skills/\$n || exit 1; done && echo ok"
 else
   if docker run --rm --entrypoint="" "$IMAGE" sh -c "command -v bun" >/dev/null 2>&1; then
     fail "bun should NOT be in base image but was found"
@@ -261,6 +240,57 @@ else
   fi
 fi
 
+echo
+echo "-- Image-baked fallback skills + harness instruction --"
+# Baked under /usr/local/share (base image, both variants). entrypoint-user.sh
+# symlinks these into ~/.agents/skills/ and ~/.config/opencode/instructions/ on
+# container start; assert the SOURCE here (smoke runs with --entrypoint="").
+run "baked opencode-devbox-environment skill" \
+    "test -f /usr/local/share/opencode-devbox/skills/opencode-devbox-environment/SKILL.md && echo ok"
+run "baked mempalace fallback skill" \
+    "test -f /usr/local/share/opencode-devbox/skills/mempalace/SKILL.md && echo ok"
+run "baked harness instruction (opencode-devbox.md)" \
+    "test -f /usr/local/share/opencode-devbox/instructions/opencode-devbox.md && echo ok"
+
+echo
+echo "-- Build provenance (manifest + OCI labels) --"
+run "/etc/opencode-devbox/build-manifest.json present" \
+    "test -f /etc/opencode-devbox/build-manifest.json"
+run_expect "manifest records opencode component" \
+    "cat /etc/opencode-devbox/build-manifest.json" '"opencode"'
+run_expect "manifest records opencode_version" \
+    "cat /etc/opencode-devbox/build-manifest.json" '"opencode_version"'
+run_expect "manifest records mempalace-toolkit component" \
+    "cat /etc/opencode-devbox/build-manifest.json" '"mempalace-toolkit"'
+run_expect "manifest records release_tag" \
+    "cat /etc/opencode-devbox/build-manifest.json" '"release_tag"'
+# The manifest is only useful if something can READ it. v2.9.0 added
+# opencode-devbox-version as that reader (and entrypoint-user.sh prints it at
+# container start), so assert the command itself works in all three modes —
+# otherwise the manifest stays an invisible artifact, which is what it was for
+# every release before this one.
+run_expect "opencode-devbox-version --json emits the manifest" \
+    "opencode-devbox-version --json" '"release_tag"'
+run "opencode-devbox-version --quiet" "opencode-devbox-version --quiet"
+run "opencode-devbox-version (human)" "opencode-devbox-version | head -1"
+# Every resolved component must be a real value, never the 'unknown'
+# sentinel that rev()/version lookups emit on failure. (oh-my-opencode-slim
+# is JSON null in the base variant — that is expected, not 'unknown'.)
+run "manifest has no unresolved ('unknown') components" \
+    "! grep -q '\"unknown\"' /etc/opencode-devbox/build-manifest.json"
+if [ "$VARIANT" = "omos" ]; then
+  run "manifest omos component is resolved (not null) in omos variant" \
+      "! grep -q '\"oh-my-opencode-slim\": null' /etc/opencode-devbox/build-manifest.json"
+fi
+# OCI labels live in the image config, not the container fs — inspect them
+# from the host docker rather than via `docker run`.
+LBL=$(docker inspect --format '{{ index .Config.Labels "se.jordbo.opencode-devbox.opencode-version" }}' "$IMAGE" 2>/dev/null || true)
+if [ -n "$LBL" ] && [ "$LBL" != "<no value>" ]; then
+  pass "OCI label se.jordbo.opencode-devbox.opencode-version=$LBL"
+else
+  fail "OCI label se.jordbo.opencode-devbox.opencode-version missing or empty"
+fi
+
 echo
 echo "-- Entrypoint behaviour --"
 
@@ -342,6 +372,37 @@ if docker run --rm \
 else
   fail "$label: existing config was modified!"
 fi
+
+# Proposed-config side-channel: when a config already exists, a NEWER default
+# config is surfaced as a NON-loaded opencode.jsonc.proposed (write-on-diff,
+# removed once the live config matches). The live config is never touched.
+label="generate-config writes .proposed only when config differs"
+if docker run --rm \
+    -e OPENCODE_PROVIDER=anthropic \
+    -e HOME=/tmp/home \
+    --entrypoint="" \
+    "$IMAGE" sh -c '
+      set -e
+      d=/tmp/home/.config/opencode
+      mkdir -p "$d"
+      gc=/usr/local/lib/opencode-devbox/generate-config.py
+      # (a) differing existing config → proposed written, live NOT clobbered
+      printf "{\n  \"model\": \"old/model\"\n}\n" > "$d/opencode.jsonc"
+      python3 "$gc" 2>/dev/null
+      test -f "$d/opencode.jsonc.proposed"
+      grep -q "old/model" "$d/opencode.jsonc"
+      # (b) live matches defaults + stale proposed present → proposed removed
+      rm -f "$d/opencode.jsonc" "$d/opencode.jsonc.proposed"
+      python3 "$gc" 2>/dev/null
+      cp "$d/opencode.jsonc" "$d/opencode.jsonc.proposed"
+      python3 "$gc" 2>/dev/null
+      test ! -f "$d/opencode.jsonc.proposed"
+      echo ok
+    ' 2>/dev/null | grep -q ok; then
+  pass "$label"
+else
+  fail "$label: proposed-config behaviour incorrect"
+fi
 rm -rf "$tmp"
 
 echo
@@ -350,29 +411,40 @@ SIZE_BYTES=$(docker image inspect --format='{{.Size}}' "$IMAGE")
 SIZE_MB=$((SIZE_BYTES / 1024 / 1024))
 echo "  Uncompressed size: ${SIZE_MB} MB"
 
-# Thresholds (uncompressed): base 2500 MB, omos 3300 MB, with-pi adds ~150 MB.
+# Thresholds (uncompressed): base 2600 MB, omos 3300 MB.
 # omos bumped 3000→3200 on v1.14.31c — mempalace-toolkit bake-in pushed the
 # baseline; bumped 3200→3300 on v1.15.0 — opencode 1.15.0 came in at
 # 3206 MB, leaving zero headroom for routine apt-get upgrade drift.
-# omos-with-pi bumped 3400→3500 on v1.15.0 alongside the omos bump.
-# omos-with-pi bumped 3500→3700 on v1.15.4b — omos+pi compounded as both
-# upstream packages grew (opencode 1.15.0→1.15.4, pi 0.74.0→0.75.3) and
-# the variant landed just over 3500 in v1.15.4's smoke.
-# with-pi 2700→2900 and omos-with-pi 3700→3900: baking pi-fork +
-# pi-observational-memory node_modules into /opt (fork pulls its
-# @earendil-works peer deps, ~150 MB) adds to both pi-bearing variants.
 # base 2500→2600 on v1.15.13c — base crept to 2506 MB (LAN-access script +
 # updated entrypoint + routine apt-get upgrade drift), tripping the
 # deliberately zero-headroom 2500 ceiling and skipping promote-base-latest.
 # omos variant to ~3.1 GB. Functional smoke checks all pass; this is a
 # guardrail, not a performance limit.
-THRESHOLD=2600
-[ "$VARIANT" = "omos" ] && THRESHOLD=3300
-[ "$VARIANT" = "with-pi" ] && THRESHOLD=2900
-[ "$VARIANT" = "omos-with-pi" ] && THRESHOLD=3900
-# pi-only = with-pi minus opencode (its platform binary is ~145 MB), so it
-# lands a bit under base. Threshold 2750 leaves the same headroom pattern.
-[ "$VARIANT" = "pi-only" ] && THRESHOLD=2750
+# v1.16.2: all thresholds bumped +150 MB preemptively ahead of the combined
+# opencode 1.15.13->1.16.2 (minor) + pi 0.78.1->0.79.0 (minor) bump. Both
+# base (2506/2600) and omos (3206/3300) were sitting on ~94 MB headroom and
+# a minor opencode bump has tripped them before (v1.15.0 omos). Restoring
+# ~250 MB headroom avoids a partial-publish + letter-suffix recovery cycle.
+# CI's smoke size print + resolved-versions table records the actual landed
+# sizes; tighten later if they come in low.
+# v2.6.0: bumped +200 MB (2750->2950 base, 3450->3650 omos). The pi-devbox
+# CLI-parity additions (pandoc ~150 MB + graphviz ~40 MB + tealdeer ~5 MB, all
+# on the base layer) consumed almost all of the ~250 MB headroom v1.16.2 had
+# restored — v2.6.0 passed 2750/3450 but with a slim margin. Restoring ~250 MB
+# headroom keeps the guardrail catching *runaway* growth (accidental texlive/
+# chrome bake-in) rather than tripping on routine apt drift or a minor opencode
+# bump. smoke still prints the actual landed size each run; tighten if low.
+# v2.9.0: bumped +650 MB (2950->3600 base, 3650->4300 omos) for agent-browser +
+# Playwright Chromium on the BASE layer (~625 MB after dropping the redundant
+# chromium_headless_shell build). Sized to keep roughly the same ~250 MB of
+# headroom the previous thresholds had, so the guardrail still catches *runaway*
+# growth (an accidental texlive or a second browser) rather than tripping on
+# routine apt drift. NOTE: do NOT copy pi-devbox's threshold number across — it
+# sums `docker history` while this script uses `docker image inspect .Size`, so
+# the two are not directly comparable. smoke prints the actual landed size every
+# run; tighten these if they come in low.
+THRESHOLD=3600
+[ "$VARIANT" = "omos" ] && THRESHOLD=4300
 if [ "$SIZE_MB" -gt "$THRESHOLD" ]; then
   fail "image size ${SIZE_MB} MB exceeds threshold ${THRESHOLD} MB for variant=$VARIANT"
 else

ssh-lan.conf.example

@@ -0,0 +1,45 @@
+# ssh-lan.conf.example — host-owned LAN-peer jump overrides for opencode-devbox
+# ============================================================================
+# WHAT THIS IS
+#   On a VM-backed host (macOS OrbStack / Docker Desktop) the container can't
+#   reach the host's LAN directly; it tunnels through the host via the `host`
+#   SSH jump that the entrypoint sets up (see the README "Reaching your LAN"
+#   section). To reach your LAN peers *by name*, they need `ProxyJump host`.
+#
+# WHY NOT JUST EDIT ~/.ssh/config?
+#   The host itself reaches those peers DIRECTLY — adding `ProxyJump host`
+#   there would break the host's own access (and ~/.ssh is mounted read-only
+#   into the container anyway). So container-only jump overrides live HERE.
+#
+# HOW IT'S WIRED
+#   If this file exists at  ~/.config/devbox-shell/ssh-lan.conf  on the host
+#   (the same bind-mounted devbox-shell bridge dir used for shared aliases),
+#   the generated ~/.ssh-local/config Includes it BEFORE your ~/.ssh/config.
+#   SSH's first-value-wins rule means ProxyJump is taken from here, while
+#   HostName / User / IdentityFile are inherited from the matching block in
+#   your ~/.ssh/config. So you only list the names + the jump — nothing else.
+#
+# SETUP
+#   1. Copy to your host:  cp ssh-lan.conf.example ~/.config/devbox-shell/ssh-lan.conf
+#   2. Bind-mount ~/.config/devbox-shell into the container (most setups
+#      already do this for shared shell aliases).
+#   3. List the host aliases (as named in your ~/.ssh/config) that should be
+#      reached through the host jump.
+#   4. Restart the container, then:  dssh <name>
+#
+# NOTE: these are facts about ONE host's LAN. A roaming laptop sees different
+# networks — keep this per-host, never in the image. For ad-hoc private IPs on
+# whatever LAN you're currently on, prefer DEVBOX_LAN_AUTOJUMP_PRIVATE=1
+# instead of naming every peer.
+
+# Example — names must match Host blocks already defined in your ~/.ssh/config:
+Host pve pve-2 pbs-vm my-nas
+    ProxyJump host
+
+# You can also give a peer its own settings here if it isn't in ~/.ssh/config
+# at all (then specify everything, not just ProxyJump):
+# Host lab-box
+#     HostName 192.168.1.77
+#     User admin
+#     IdentityFile ~/.ssh/id_ed25519
+#     ProxyJump host