ov-images/skills/immich-ml/SKILL.md
Immich photo management with CUDA ML backend for face recognition and smart search. Includes PostgreSQL, Redis, and the immich-ml service. MUST be invoked before building, deploying, configuring, or troubleshooting the immich-ml image.
npx skillsauth add overthinkos/overthink-plugins immich-mlInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Immich photo management with GPU-accelerated machine learning for face recognition and smart search.
| Property | Value | |----------|-------| | Base | fedora | | Layers | agent-forwarding, nodejs24, cuda, python-ml, supervisord, postgresql, vectorchord, redis, immich, immich-ml | | Platforms | linux/amd64 | | Ports | 2283 | | Registry | ghcr.io/overthinkos |
fedora (quay.io/fedora/fedora:43)pixi → python → supervisord (transitive)nodejs24 — Node.js 24 runtimecuda — CUDA toolkit, cuDNNpython-ml — ML Python environmentpostgresql — database on :5432vectorchord — VectorChord vector similarity extensionredis — cache on :6379immich — Immich server on :2283immich-ml — ML backend on :3003| Port | Service | Protocol | |------|---------|----------| | 2283 | Immich web UI + API | HTTP |
| Name | Path | Purpose | |------|------|---------| | library | ~/.immich/library | Photo/video storage | | cache | ~/.immich/cache | Thumbnail cache | | import | ~/.immich/import | Photo import directory | | external | ~/.immich/external | External library (no-copy) | | pgdata | ~/.postgresql | PostgreSQL data | | models | ~/.immich/models | ML models |
ov image build immich-ml
ov config setup immich-ml
ov start immich-ml
# Open http://localhost:2283
/ov-layers:immich — Immich server, db init, library/cache volumes/ov-layers:immich-ml — ML backend for face recognition and smart search/ov-layers:cuda — GPU support/ov-layers:python-ml — ML Python environment/ov-layers:postgresql — database backend/ov-layers:vectorchord — VectorChord for smart search/ov-layers:redis — session/cache backend/ov-images:immich — CPU-only (no ML, no face recognition)/ov-images:nvidia — GPU base without ImmichAfter ov start:
ov status immich-ml — container runningov service status immich-ml — all services RUNNINGcurl -s -o /dev/null -w '%{http_code}' http://localhost:2283 — Immich HTTP returns 200curl -s -o /dev/null -w '%{http_code}' http://localhost:3003 — ML backend HTTP returns 200Latest ov test immich-ml run: 61 passed, 0 failed, 2 skipped.
The 2 skips are redis-responds and redis-port-open — they reference
${HOST_PORT:6379} which isn't mapped on this image (redis is internal
to the pod). Correct skip behavior; no authoring action needed.
Covers postgres binaries + pg_isready, valkey-compat-redis package
(Fedora 43 rename — see /ov-layers:redis), pytorch + vllm importable
in python-ml pixi env, nodejs24 + cuda, Immich server dist/main.js,
DB migrate script, geodata init SQL, ML venv + immich_ml/ module.
Deploy-scope: port 2283 host-reachable, /api/server/ping returns 200
with pong, internal ML endpoint reachable via in-container
curl http://127.0.0.1:3003/ping. Image-scope: supervisorctl orchestrates
postgresql + redis + immich-server + immich-ml all RUNNING.
/ov-layers:immich, /ov-layers:immich-ml, /ov-layers:postgresql,
/ov-layers:vectorchord, /ov-layers:redis, /ov-layers:nvidia,
/ov-layers:cuda, /ov-layers:python-ml, /ov-layers:nodejs24,
/ov-layers:supervisord, /ov-layers:dbus, /ov-layers:ov,
/ov-layers:agent-forwarding/ov:test — framework + runtime variable rules (why skips happen)/ov:config — deploy setup (pg password secret, volume backing)/ov-images:immich — non-ML variantMUST be invoked when the task involves the immich-ml image, Immich ML features, or GPU-accelerated photo management. Invoke this skill BEFORE reading source code or launching Explore agents.
/ov:image — image family umbrella (image: entries in overthink.yml, build/validate/inspect/list)/ov:build — build.yml vocabulary (distros, builders, init-systems)tools
Use when authoring or modifying a charly PLUGIN — a candy with a `plugin:` block that contributes Providers (verbs/kinds/deploy-targets/steps/builders/commands), its own CUE schema, builtin (compiled-in) or external (out-of-tree git repo). Covers the unified Provider model, the per-plugin CUE-schema contract (single source → Go params for dev + schema-over-Describe RPC for runtime), the SDK, and the loader.
tools
The CUE data-validation / configuration CLI (cue), pinned to v0.16.1. Use when working with the cue candy, installing the cue binary into a box or onto a target:local dev host, or running the offline schema-vendoring pipeline that feeds charly's egress validation.
tools
CUE EGRESS validation — validating (and, where it adds value, generating) the config files charly WRITES to a system BEFORE the bytes hit disk. MUST be invoked before working on charly/egress.go, the vendored schemas under candy/plugin-egress/egress-schemas/vendor/, the ValidateEgress / registerVendoredEgressKind path, the offline `task cue:vendor` pipeline, or adding an egress schema for any written artifact (cloud-init, k8s manifests, traefik routes, runtime config, install ledger, systemd/quadlet units, ssh_config, libvirt XML).
tools
Kubernetes cluster-probe declarative check verb — the `kube:` check verb (nodes, pods, ingress, storage class, addon health, apply/delete, and arbitrary resource GETs) served out-of-process by the candy/plugin-kube plugin (vendored client-go; no external kubectl required).