Minimal helper to audit skill.md-style instructions for supply-chain risks.
Initial release of the security-audit skill. - Provides heuristic scanning of skill.md instructions for supply-chain risks. - Detects exfiltration patterns such as HTTP POST requests, unknown domain curl commands, ~/.env reads, and credential keywords. - Summarizes detected filesystem and network accesses as a permission manifest reminder. - Generates a safe report in markdown format with an assigned risk level. - Simple command-line usage via Python.