Review catalog findings.
Inspect descriptions, input types, response schemas, and documented authentication. Find metadata gaps that make integration harder.
API TESTING
Bring catalog checks and hands-on tool testing into the same workflow. Catch unclear inputs and surprising responses before connecting an agent.
// Enter pay_1042 for a successful response.
// Try another ID to inspect an error.01 / CHECK
Static quality checks and live tool calls answer different questions. Sutra gives you a place to work through both.
Inspect descriptions, input types, response schemas, and documented authentication. Find metadata gaps that make integration harder.
Use the playground to list available tools and call them with explicit inputs. Read the response before handing the workflow to an agent.
Export a collection from your contract and continue testing with the tools your team already uses. Keep the exported surface aligned with the intended audience.
02 / INVESTIGATE
Work outward from the input. Is the identifier valid? Is the tool exposed? Can the upstream service authorize this operation?
A playground call can reach a real upstream service. Use test credentials and non-production data, especially for tools that create, update, or delete records.
This page describes catalog quality checks, MCP playground calls, and Postman exports. It does not promise a built-in scheduled or chaos-testing runner.
No. Metadata checks reveal specific issues in a catalog. They do not replace authorization testing, penetration testing, or a review of the upstream service.
Yes. In the example above, keep pay_1042 for success or enter a different identifier for a simulated not-found response. The example is entirely local to this page.