Use APIContext with a coding agent
The primary source is the complete approved OpenAPI JSON. It covers API v2 and v3 and includes operations, parameters, schemas, error responses, and security requirements. The smaller curated downloads are incomplete subsets.
- Fetch and parse
https://docs.apicontext.com/openapi/api3-full.json; do not scrape generated HTML as the contract. - Choose a read-only GET initially, such as
GET /api/3/monitors/api/or the legacyGET /api/2/calls/. Use the operation ID and method/path in the spec, not guesses. - Combine the declared server
https://client.apimetrics.iowith the full versioned path. Do not prepend /api/2 or /api/3 twice. - Read the operation's security requirement and
components.securitySchemes. The full contract currently declares OAuth2 and X-Api-Key; some compatibility references use an Authorization header. Do not invent or hardcode credentials. - Read required parameters and success/error schemas, paginate where necessary, and test with a user-authorized nonproduction project.
Example task for a coding agent
Build a CLI that lists my API monitors using the complete API v3 contract. Read credentials from environment variables; do not log them or store them in source. Validate required parameters, support pagination, and handle 401, 403, 429, and transient failures. Do not modify or delete any remote data.
The account/project identifiers and credentials must be supplied by an authorized user. A coding agent cannot bypass API access controls.