Agent Skills: Clous API Integration

Use when building or reviewing direct Clous API v1 integrations, including responses, object CRUD, API keys, webhooks, tool endpoints, retries, pagination, and structured error handling.

UncategorizedID: clous-ai/agents/api-integration

Repository

clous-aiLicense: NOASSERTION

Install this agent skill to your local

pnpm dlx add-skill https://github.com/clous-ai/plugins/tree/HEAD/clous-development/skills/api-integration

Skill Files

Browse the full folder contents for api-integration.

Download Skill

Loading file tree…

clous-development/skills/api-integration/SKILL.md

Skill Metadata

Name
api-integration
Description
Use when building or reviewing direct Clous API v1 integrations, including responses, object CRUD, API keys, webhooks, tool endpoints, retries, pagination, and structured error handling.

Clous API Integration

Use the Clous API as the source of truth for production integrations. The primary base URL is https://api.clous.app/v1.

Workflow

  1. Discover capabilities with GET /v1/info/.
  2. Choose auth: API key for server-to-server, OAuth token for delegated user context.
  3. Model requests against stable resources: responses, objects, api-keys, webhooks, and tools.
  4. Implement structured error handling for 401, 403, 404, 429, and 5xx responses.
  5. Validate with the Python and TypeScript SDKs before handoff.

Endpoints

  • GET /v1/info/: API discovery.
  • POST /v1/responses/, GET /v1/responses/{id}/: agent runs.
  • GET /v1/objects/types/: supported object resources.
  • GET|POST /v1/objects/{type}/: object collection operations.
  • GET|PATCH|DELETE /v1/objects/{type}/{id}/: object detail operations.
  • GET|POST /v1/webhooks/, GET /v1/webhooks/events/: webhook management.
  • POST /v1/tools/hr-knowledge-retrieve/, POST /v1/tools/clous-support-docs-retrieve/, POST /v1/tools/systemlog-errors-retrieve/: worker tool endpoints.

Guardrails

  • Do not log bearer tokens, OAuth codes, refresh tokens, webhook secrets, or employee payloads.
  • Prefer idempotent retries only for GET and explicitly safe POST calls.
  • For mutations, include a human-readable intent, actor, and correlation ID when the host supports headers.