Skip to content

http-contract.md

Latest commit

 

History

History
40 lines (29 loc) · 1.54 KB

http-contract.md

File metadata and controls

40 lines (29 loc) · 1.54 KB
Raw

ClawPro HTTP backend

Part of the TeamAI CLI Usage Guide.

ClawPro implements these three endpoints for teamai init --http <baseUrl>. Send Authorization: Bearer <api-key>.

End users only need teamai init --http, described in Member Onboarding.

Endpoint Method Purpose
{baseUrl}/api/local-agent/report POST Session start: register the agent and installed skills
{baseUrl}/api/local-agent/sync POST Report status and return commands to run
{baseUrl}/api/local-agent/commands/ack POST Acknowledge one command: { id, status, error }

sync returns the commands:

{
  "ok": true,
  "commands": [{ "id": 1, "type": "install_skill", "skill_slug": "x", "skill_version": "1.0.0", "download_url": "https://signed-url/..." }]
}

Push models with apply_model_config. cmd is a JSON string. {"models":[...]} replaces the whole set.

{ "id": 16, "type": "apply_model_config",
  "cmd": "{\"models\":[{\"provider\":\"openai\",\"model_id\":\"gpt-4o\",\"name\":\"GPT-4o\",\"base_url\":\"https://proxy.example.com/v1\",\"api_key\":\"<ProxyToken>\",\"max_tokens\":4096,\"context_window\":128000}]}" }

Installed models are reported through report. Omit models when there is nothing to report. api_key is not sent back.

{ "agent_type": "codebuddy", "local_agent_id": "...",
  "user_level": { "models": [
    { "provider": "tokenhub", "model_id": "gpt-4o", "name": "GPT-4o", "source": "enterprise" }
  ] } }