AI agents can write code, but they can't write, run, and fix API tests in a closed loop — until now. Glubean gives Claude the SDK patterns, MCP tools, and structured failure data to autonomously generate typed TypeScript API/Graphql/Grpc/Browser tests, execute them, read structured failures, and iterate until green.
No more Postman collections. No more manual HTTP testing. One prompt, and the agent handles the rest.
Created by @glubean. Full skill repo: https://github.com/glubean/skill
Write CRUD tests for the GitHub Issues API, run them, and fix any failures
The AI bootstraps the project, configures auth (with user confirmation), generates typed tests, runs them via MCP, reads structured failures, and iterates — all from a single prompt.
Tested with Claude Code via npx skills add glubean/skill.