specification: API Commons Agent Skills
specificationVersion: '0.1'
provider: AIMLAPI
providerId: aimlapi
generated: '2026-08-30'
method: generated
source: >-
  Authored from openapi/aimlapi-inference-openapi.yml (the provider's own
  published contract at https://api.aimlapi.com/docs-yaml), the per-endpoint
  OpenAPI fragments embedded in the AIMLAPI service-endpoint documentation, and
  the conventions, errors, lifecycle and scopes artifacts in this repo.
description: >-
  Packaged Agent Skills for AI/ML API. Every operationId referenced is present in
  the published contract or in a documented per-endpoint fragment; none was
  invented. A provider-published skills file or AGENTS.md was searched for and
  does not exist.
searched_for_provider_published:
  aimlapi.com/AGENTS.md: not found
  docs.aimlapi.com skills: not found
  note: >-
    The provider does publish an llms.txt at https://aimlapi.com/llms.txt and a
    GitBook "Agent Instructions" footer on every docs page describing a
    ?ask= query interface, but no packaged Agent Skills.
skills:
- file: aimlapi-choose-and-call-a-model.md
  name: aimlapi-choose-and-call-a-model
  summary: >-
    Resolve a model from the live catalogue and check the deprecation feed before
    calling it, rather than hardcoding an id that may already be withdrawn.
  operations:
  - GET /v1/models
  - GET /v1/models/deprecations
  - _v1_chat_completions
  - _v1_embeddings
  - _v1_images_generations
- file: aimlapi-track-what-a-call-cost.md
  name: aimlapi-track-what-a-call-cost
  summary: >-
    Read per-request cost off the response headers, correlate it with your own
    id, and reconcile against the usage logs and the billing ledger.
  operations:
  - _v1_chat_completions
  - _v2_logs
  - GET /v2/usage
  - GET /v2/billing
  - GET /v2/billing/transactions
- file: aimlapi-run-a-long-generation.md
  name: aimlapi-run-a-long-generation
  summary: >-
    Submit and poll asynchronous video, music and speech generations, run
    batches, and cancel one — the only reversal in the API.
  operations:
  - _v2_video_generations
  - _v1_stt_create
  - _v2_generate_audio
  - _v1_batches
  - _v1_batches_cancel_batch_id
- file: aimlapi-run-an-agent-safely.md
  name: aimlapi-run-an-agent-safely
  summary: >-
    Mint a model-scoped, dollar-capped API key for an agent, watch it, revoke it,
    and understand what the coarse mcp:invoke OAuth scope grants by comparison.
  operations:
  - KeysController_createKey_v1
  - KeysController_listKeys_v1
  - PATCH /v1/keys/{prefix}
  - DELETE /v1/keys/{prefix}
  - GET /v2/usage