"author:RudrenduPaul" matching MCP servers:

Exposes a run_suite tool to evaluate whether an AI agent is safe to operate internal web apps, scoring task completion and forbidden-action violations to gate CI/CD pipelines.