L’interface est localisée ; les présentations et documents des Skills sont traduits automatiquement par IA dans votre langue. Le code, les commandes, chemins et termes techniques restent en version originale, et le document de l’auteur fait foi.
Explorer les Skills
Déploiement et opérations · SKILL.md

services-and-extensions

services-and-extensions

Routes ODS service manifests, Docker Compose layering, extension catalog behavior, install/update/rollback semantics, compose security scanning, and the service registry.

Independent GitHub snapshot · Automated classification · Original documentation governs

Documentation de la Skill

GitHub

name: services-and-extensions description: "Routes ODS service manifests, Docker Compose layering, extension catalog behavior, install/update/rollback semantics, compose security scanning, and the service registry." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

Services And Extensions

Purpose

Use this sub-skill for ODS work that touches service manifests, compose overlays, extension catalog entries, extension enable/disable/install/update/rollback semantics, security-scanned compose resolution, or the service registry view of enabled services.

This sub-skill is router-like. The detailed contracts live in:

  • references/compose-and-extensions.md
  • references/service-catalog.md
  • references/troubleshooting.md
  • scripts/extension_manifest_summary.py

Covers

  • extensions/services/*/manifest.yaml and extension compose overlays
  • bundled extension library / installable catalog semantics
  • category, dependency, alias, port, GPU backend, and service.llm contracts
  • compose layering and security scan behavior
  • enable/disable/install/update/rollback state rules
  • service registry behavior and manifest-driven discovery

Route elsewhere

  • Dashboard Extensions page implementation -> dashboard-and-api
  • Exact ods enable/disable/list/audit command behavior and other host CLI details -> ops-cli-and-host-tools
  • Installer phase ordering / platform install behavior -> installers-and-platforms
  • Model selection, tier mapping, and backend capability decisions -> hardware-and-models

Fast path

  1. Run the bundled read-only summary helper on the relevant catalog root.
  2. Read references/compose-and-extensions.md for the manifest and compose contract.
  3. Read references/service-catalog.md for the current bundled service snapshot.
  4. Use references/troubleshooting.md when a manifest, overlay, or lifecycle action fails.
  5. Only then edit manifests, compose fragments, or library entries.

Working rules

  • Prefer read-only inspection first; the summary helper is safe by default and supports --help.
  • Keep ports loopback-bound unless the manifest explicitly opts into host networking.
  • Keep new service ids and aliases unique; do not shadow core service ids.
  • For LLM-consuming services, declare service.llm and default to the gateway route instead of pinning a concrete model name.
  • When a service is enabled or disabled, the compose file marker is the source of truth: compose.yaml means enabled, compose.yaml.disabled means disabled.

What to check before changing a service

  • The manifest schema version and required fields
  • The category (core, recommended, optional)
  • The compose file location and overlay naming
  • GPU backend compatibility
  • Alias collisions and dependency references
  • Security scan failure modes for any new compose content
  • Whether the service is installable from the bundled library or is only a bundled runtime service

If the task is about portal UI behavior or host command syntax, switch sub-skills instead of expanding this one.

Skills enregistrées