The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes legitimate, industry-standard security tools such as npm audit, pip-audit, govulncheck, and osv-scanner to perform repository scans. These executions are core to the skill's stated purpose and follow expected behavior for a security auditing tool.
[DATA_EXFILTRATION]: Includes defensive instructions to prevent the accidental exposure of sensitive data. It explicitly forbids the agent from printing the values of any hardcoded secrets found during the scan, requiring only references to the file path and line number.
[SAFE]: Implements a robust 'Health Engine' framework that ensures all remediations are verified against the project's existing tests and build systems. It requires explicit user approval before applying any changes and specifically instructs the agent never to perform automated commits.

absolute-audit