skill-security-auditor

Scan Python and Bash scripts for dangerous patterns like os.system, eval, and subprocess with shell by running a static security audit on the local skill directory, which outputs a machine-readable report with critical, high, and info findings.

Dependency supply-chain risk detection for AI skills is the process of scanning a skill's dependencies and filesystem access boundaries before installation to identify potential vulnerabilities and output a clear security verdict with actionable remediation steps.

Yes, you can audit a git repository for security risks because the scanner supports both local skill directories and git repositories, analyzing Python/Bash scripts, markdown content, and related assets for dangerous patterns and data exfiltration.

Yes, the security audit detects potential data exfiltration by statically scanning Python and Bash scripts for dangerous patterns like os.system, eval, and subprocess with shell, generating a machine-readable report with critical findings.

A limitation of static scanning for prompt-injection in markdown files is that it identifies suspicious instructions altering behavior or bypassing safety, but may require manual review to confirm contextual threats within the generated security report.