Show HN: Preflight-agent – five checks your AI-agent setup needs before it runs
Show HN: Preflight-agent – five checks your AI-agent setup needs before it runs
preflight-agent is a Python 3.10+ command-line tool that runs five read-only static audits on an AI-agent setup in one scan: clone-guard (auto-execution surfaces), secret-guard (inline secrets in agent/MCP JSON configs), hook-doctor (dead or non-blocking Claude Code hooks), fleet-audit (agent-definition collisions and grants) and skill-sprawl (skill duplicates, collisions and bloat). It prints a combined receipt, supports --json, and returns exit codes 0 (clean), 1 (findings) or 2 (error); it is on PyPI since v1.0.1 and uses only the standard library with no network access. The suite merges five previously standalone auditors whose engines are unchanged, and the author notes that a CLEAN verdict covers only what the five modules check, not a guarantee of safety.
Source: Hacker News AI · github.com