generated: '2026-08-16'
method: generated
source: >-
  openapi/furiosa-predict-v2.yaml, openapi/furiosa-model-repository-v2.yaml,
  https://developer.furiosa.ai/latest/en/furiosa_llm/furiosa-llm-serve.html,
  https://developer.furiosa.ai/latest/en/furiosa_llm/responses-api.html
note: >-
  Packaged operating instructions for the two Furiosa HTTP surfaces, grounded in real published
  operationIds (Model Server) and real documented method+path pairs (Furiosa-LLM, which has no
  published OpenAPI). Nothing here is invented.
  Separately: FuriosaAI does publish Agent Skills of its own at
  https://github.com/furiosa-ai/agent_skills (commit-msg, pr-create, pr-restructure, pr-update,
  prepare-docs, write-docs, update-docs, distributed as a Claude Code plugin marketplace). Those
  are FuriosaAI's INTERNAL engineering-workflow skills — they automate its own docs and PR
  process and say nothing about calling FuriosaAI's APIs — so they are recorded here as a
  finding rather than copied into this directory.
provider_published_skills:
  repo: https://github.com/furiosa-ai/agent_skills
  kind: internal engineering workflow
  skills: [commit-msg, pr-create, pr-restructure, pr-update, prepare-docs, write-docs, update-docs]
  covers_provider_api: false
skills:
  - file: furiosa-serve-a-model.md
    name: furiosa-serve-a-model
    api: furiosa-model-repository-v2
    description: Load a model into a Furiosa Model Server and verify readiness before inference.
    operations:
      - post-v2-repository-index
      - post-v2-repository-models-$-MODEL_NAME-load
      - get-v2-models-$-modelName-versions-$-modelVersion-ready
      - get-v2-models-$-modelName-versions-$-modelVersion
      - get-v2-health-ready
  - file: furiosa-run-inference.md
    name: furiosa-run-inference
    api: furiosa-server-predict-v2
    description: Build and send a KServe v2 inference request from the model's own tensor metadata.
    operations:
      - get-v2
      - get-v2-models-$-modelName-versions-$-modelVersion
      - post-v2-models-$-MODEL_NAME-versions-$-MODEL_VERSION-infer
  - file: furiosa-llm-openai-client.md
    name: furiosa-llm-openai-client
    api: furiosa-llm-openai-server
    description: Call the Furiosa-LLM OpenAI-compatible server, including every documented deviation from OpenAI and OpenResponses.
    operations:
      - GET /v1/models
      - POST /v1/chat/completions
      - POST /v1/completions
      - POST /v1/responses
      - POST /v1/embeddings
      - GET /metrics
skill_count: 3