Merge pull request #2667 from alexbelgium/copilot/fix-claude-md-documentation

docs(CLAUDE.md): fix three inaccuracies in agent guidance
This commit is contained in:
Alexandre
2026-04-22 07:27:01 +02:00
committed by GitHub

View File

@@ -8,24 +8,25 @@ This is a Home Assistant add-on repository containing 120+ Docker-based add-ons
## Add-On Directory Structure
Every add-on follows this layout:
Most add-ons follow this common layout, though exceptions exist (e.g. some archived add-ons use `config.json` instead of `config.yaml`, some add-ons have `build.yaml` instead of `build.json` or no build file at all, and not every add-on includes a `rootfs/` tree):
```
addon_name/
├── config.yaml # HA add-on metadata, schema, ports, maps
├── build.json # Base Docker images per architecture
├── build.json # Base Docker images per architecture (may be build.yaml, or absent)
├── Dockerfile # Multi-stage build (always uses shared .templates/ scripts)
├── updater.json # Upstream release tracking (used by addons_updater)
├── updater.json # Upstream release tracking; required to enable automatic updates
├── CHANGELOG.md # Required; must be updated on every PR
└── rootfs/
└── rootfs/ # Optional; absent in some add-ons
└── etc/
├── cont-init.d/ # S6-overlay init scripts (numbered, run in order)
└── services.d/ # S6-overlay supervised services
└── services.d/ # S6-overlay supervised services (some add-ons use
# s6-overlay v3 layout at etc/s6-overlay/s6-rc.d/ instead)
```
## Dockerfile Convention
All Dockerfiles follow a strict 6-section pattern:
Most Dockerfiles follow this 6-section pattern (some add-ons deviate slightly, e.g. using a pinned upstream image directly instead of `ARG BUILD_FROM`):
1. **Build Image** `ARG BUILD_FROM` + `FROM ${BUILD_FROM}`
2. **Modify Image** S6 env vars, LSIO modifications via `ha_lsio.sh`
@@ -37,11 +38,11 @@ All Dockerfiles follow a strict 6-section pattern:
Shared build-time scripts are pulled from `.templates/` at build time:
- `ha_automodules.sh` Downloads module scripts listed in `ARG MODULES=`
- `ha_autoapps.sh` Installs packages listed in `ENV PACKAGES=`
- `ha_entrypoint.sh` S6 stage-2 hook; converts `options.json` to env vars
- `ha_entrypoint.sh` S6 stage-2 hook; launches the cont-init stack at container start
- `ha_lsio.sh` Patches LinuxServer.io base images for HA compatibility
- `bashio-standalone.sh` Bashio library for scripts outside Supervisor context
The `ARG MODULES=` line lists template scripts to download at build time (e.g., `00-banner.sh 01-custom_script.sh 00-smb_mounts.sh`). Available modules in `.templates/`:
The `ARG MODULES=` line lists template scripts to download at build time (e.g., `00-banner.sh 01-custom_script.sh 00-smb_mounts.sh`). Commonly-used modules in `.templates/` (not exhaustive):
- `00-global_var.sh` Initialize global env vars from HA options
- `00-local_mounts.sh` Mount local disks (localdisks option)
- `00-smb_mounts.sh` SMB/CIFS network mount support
@@ -59,7 +60,7 @@ Key fields in every add-on's `config.yaml`:
```yaml
arch: [aarch64, amd64]
image: ghcr.io/alexbelgium/{slug}-{arch}
version: "X.Y.Z-N" # upstream version + patch suffix
version: "X.Y.Z" # upstream version (format varies; see Versioning section)
ingress: true/false
ingress_port: 8000
map:
@@ -78,11 +79,19 @@ schema:
localdisks: str? # Local disk mounts
```
The `env_vars` schema key enables the `ha_entrypoint.sh` passthrough mechanism, which converts all options in `/data/options.json` to environment variables at runtime.
The `env_vars` schema key enables the env-var passthrough mechanism. At runtime the `00-global_var.sh` cont-init module reads `/data/options.json` and exports each key as an environment variable (writing to `/.env` and `/etc/environment`). `ha_entrypoint.sh` is the S6 stage-2 hook that launches the cont-init stack but does not itself perform the JSON-to-env conversion.
## Versioning
Add-on versions use the format `X.Y.Z-N` where `X.Y.Z` is the upstream application version and `-N` is a patch counter. Both `config.yaml` and `build.json` (via `BUILD_UPSTREAM` arg) must be updated together. The `updater.json` file tracks which upstream source/repo to monitor and records the last seen version.
Add-on versions in `config.yaml` closely follow the upstream release tag and do not conform to a single fixed format. Common patterns include:
- `X.Y.Z` plain upstream semver (e.g. `0.137.0`)
- `X.Y.Z-N` upstream version with a local patch counter (e.g. `0.6.26-2`)
- LSIO-style tags (e.g. `1.43.1.10611-1e34174b1-ls301`)
- Date-based versions (e.g. `2026.02.28`)
- Nightly builds (e.g. `nightly-20260321-397`)
When an upstream version is bumped, update `version` in `config.yaml`. If the add-on's `Dockerfile` contains an `ARG BUILD_UPSTREAM` line, update that value too — it is the canonical place that records the upstream version at build time (it is **not** stored in `build.json`/`build.yaml`). Some add-ons do not use `BUILD_UPSTREAM` at all. The `updater.json` file tracks which upstream source/repo to monitor and records the last seen version.
## updater.json Format
@@ -131,8 +140,9 @@ Adding `[nobuild]` anywhere in a commit message skips the builder workflow.
## S6-Overlay Init Script Naming
Scripts in `rootfs/etc/cont-init.d/` run in lexicographic order. The conventional numbering:
Scripts in `rootfs/etc/cont-init.d/` run in lexicographic order. Common numbering conventions (many add-ons use other prefixes too):
- `20-*` Directory/folder setup
- `32-*` Nginx ingress configuration (e.g. `32-nginx_ingress.sh`)
- `80-*` Application configuration
- `90-*` Ingress / nginx setup
- `90-*` Misc pre-startup tasks (ssl, vpn, custom run)
- `99-*` Final startup / launch