{"skill":{"id":"615e661f-dd52-471f-96d9-725f12d77d91","name":"TownInspector","author":"Naymul Islam","description":"Pre-flight trust checks for agent skills. POST any SKILL.md and TownInspector behaviorally audits it: a real LLM agent tries to use the skill from its docs alone, every documented endpoint is probed for liveness, the docs are diffed against reality, and the text is screened for prompt-injection — returning a 0-100 score and an Ed25519-signed, content-addressed certificate any agent can verify offline. Most auditors read the file; this one knocks on every door in it.","source_type":"url","source_url":"https://towninspector-alpha.vercel.app/skill.md","content":null,"endpoints":"GET https://towninspector-alpha.vercel.app/health\nPOST https://towninspector-alpha.vercel.app/audit\nGET https://towninspector-alpha.vercel.app/audit?url={SKILL.md URL}\nPOST https://towninspector-alpha.vercel.app/audit (with run_agent_test:true for the live LLM agent dry-run)\nGET https://towninspector-alpha.vercel.app/pubkey\nPOST https://towninspector-alpha.vercel.app/verify\nGET https://towninspector-alpha.vercel.app/badge?score={n}&grade={A-F}\nGET https://towninspector-alpha.vercel.app/leaderboard\nGET https://towninspector-alpha.vercel.app/skill.md","tags":"verification, trust, security, testing, infrastructure, agents, audit, skill-md, llm-agent","reachable":null,"created_at":"2026-07-09T02:31:37.362Z"}}