fetchsandbox/mcp ? reverse-engineered prompt

Reverse engineered prompt

Build me an MCP server that lets my AI agent turn any OpenAPI spec into a real sandbox it can actually use from my IDE.

I want to paste in an OpenAPI URL or the spec content, have it import the API, show me the available workflows, and let the agent run one end to end with realistic, schema checked responses. It should work with MCP compatible tools like Cursor or Claude Code through stdio, and it should be easy to start with npx instead of a global install.

For the first version, make sure it can handle a few common APIs well, and if it sees a spec it does not know, it should still infer useful create and verify style workflows. I also want a simple way to inspect the imported spec, see the run trace, and get a shareable link for the result. If you need to look up current docs online, go ahead.

Are you gonna build this?

make sure you review the code using coderabbit

Try freeSponsored — opens CodeRabbit in a new tab