generated: '2026-08-16'
method: generated
source: >-
  Generated from openapi/secton-api-chat-api-openapi.yml and
  openapi/secton-api-models-api-openapi.yml, with runtime rules drawn from
  conventions/secton-api-conventions.yml, errors/secton-api-problem-types.yml,
  rate-limits/secton-api-rate-limits.yml and lifecycle/secton-api-lifecycle.yml.
  Every operationId below was grepped from the specification; none is invented.
provider_published_skills: none-found
provider_published_note: >-
  No provider-authored AGENTS.md, skills/ directory or agent instructions were found on any Secton
  host or in the sectoncorp GitHub org (0 public repositories) or the wearesecton GitLab group
  (group API 404s; only the secton-sdk project is reachable).
skill_count: 2
skills:
  - name: secton-api-chat-completion
    file: secton-api-chat-completion.md
    api: secton-api:secton-api-chat-api
    operations:
      - post-external-api-v1-chat-completions
    consequence: write
    billable: true
    idempotent: false
    summary: >-
      Generate a chat completion, blocking or streamed, with the correct auth header, the real
      error envelope, and the retry rules this API actually supports.
  - name: secton-api-list-models
    file: secton-api-list-models.md
    api: secton-api:secton-api-models-api
    operations:
      - get-external-api-v1-models
    consequence: read
    billable: false
    idempotent: true
    summary: >-
      Discover which models a key can use, and validate a credential without spending credits — the
      only safe health-check operation this API publishes.
coverage:
  openapi_operations: 2
  operations_covered: 2
  coverage_pct: 100