Files
opencode-devbox/.env.example
T
joakimp 9df126c7a9
Validate / docs-check (push) Successful in 23s
Validate / validate-base (push) Has started running
Validate / validate-omos (push) Has started running
Validate / validate-with-pi (push) Has been cancelled
Validate / validate-omos-with-pi (push) Has been cancelled
Fix: developer-writable npm prefix for pi install
NPM_CONFIG_PREFIX is now /home/developer/.pi/npm-global, with that
prefix's bin/ prepended to PATH. Without this, 'pi install npm:<pkg>'
(and any 'npm install -g') by the developer user would EACCES against
the system prefix (/usr).

The new prefix lives on the devbox-pi-config named volume, so:
  - User-installed pi packages (themes, skills, extensions) survive
    container recreate AND image rebuild, complementing pi's auto-
    restore from settings.json with one less cold-start step.
  - A user-driven 'npm install -g @mariozechner/pi-coding-agent' lands
    on the volume and wins over the baked pi via PATH order.

Build-time 'npm install -g' calls (opencode, pi, oh-my-opencode-slim)
are unaffected: the new ENVs are declared after those steps in the
Dockerfile, so the baked binaries still install to /usr at build time
and are not shadowed by the volume mount at runtime.

Verified end-to-end with a Bun-driven smoke test: as developer,
'npm install -g cowsay' inside the container succeeds, the binary
lands on PATH, and survives a fresh container against the same volume.

DOCKER_HUB.md regenerated (24997/25000 bytes, 3-byte headroom — was
138 before; future README additions to the persistence section need
to trim something else first).

Docs updated: Dockerfile inline comments, README persistence section,
AGENTS install contract, DOCKER_HUB persistence table, .env.example
notes, CHANGELOG Unreleased entry.
2026-05-09 15:41:33 +02:00

99 lines
4.8 KiB
Bash

# opencode-devbox environment configuration
# Copy this file to .env and fill in your values:
# cp .env.example .env
# ── LLM Provider ─────────────────────────────────────────────────────
# 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
# ── API Keys (set the one matching your provider) ────────────────────
# ANTHROPIC_API_KEY=
# OPENAI_API_KEY=
# GEMINI_API_KEY=
# ── AWS Bedrock (if using amazon-bedrock provider) ───────────────────
# AWS_REGION=eu-west-1
# AWS_PROFILE=default
# AWS_ACCESS_KEY_ID=
# AWS_SECRET_ACCESS_KEY=
# ── Git Configuration ────────────────────────────────────────────────
GIT_USER_NAME=
GIT_USER_EMAIL=
# ── Workspace ────────────────────────────────────────────────────────
# Path on host to mount as /workspace in the container
WORKSPACE_PATH=~/projects
# Path to SSH keys on host
SSH_KEY_PATH=~/.ssh
# ── Skillset (agent skills and instructions) ─────────────────────────
# If you have a skillset repo, the entrypoint auto-deploys skills and
# instructions on container start using relative symlinks (portable
# across host/container).
#
# Detection is automatic if the skillset lives directly at the workspace
# root (i.e. WORKSPACE_PATH/skillset → /workspace/skillset in container).
#
# If the skillset lives in a subdirectory of your workspace, set
# SKILLSET_CONTAINER_PATH to its location *inside the container*. This
# is determined by the workspace mount: whatever is at
# WORKSPACE_PATH/<subpath> on the host becomes /workspace/<subpath>
# in the container.
#
# Examples:
# Host skillset at ~/projects/skillset → already at /workspace/skillset (auto-detected, no config needed)
# Host skillset at ~/projects/tools/skillset → SKILLSET_CONTAINER_PATH=/workspace/tools/skillset
# Host skillset at ~/projects/local/skillset → SKILLSET_CONTAINER_PATH=/workspace/local/skillset
#
# Alternatively, mount the skillset repo at a dedicated path using the
# SKILLSET_PATH volume in docker-compose.yml (see comments there). In
# that case the entrypoint finds it at ~/skillset automatically.
#
# SKILLSET_CONTAINER_PATH=
# ── Locale (defaults to en_US.UTF-8) ─────────────────────────────────
# LANG=sv_SE.UTF-8
# LANGUAGE=sv_SE:sv
# LC_ALL=sv_SE.UTF-8
# ── oh-my-opencode-slim (multi-agent orchestration) ──────────────────
# 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
# ── 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)