The Agent Skills Directory

[SAFE]: The skill is a specialized diagnostic tool for security auditing and compliance. It leverages the agent's current connection to perform authorized reconnaissance and vulnerability assessments without requiring additional credentials or tokens.
[SAFE]: All referenced resources, including testing playbooks and the probe_runner.py script, are relative paths within the local repository environment, avoiding any untrusted remote code execution or external downloads.
[SAFE]: The instruction set includes strong safety guardrails, such as a 'Human Gate' policy for write operations and explicit redaction requirements for sensitive data discovered during testing.

connected-mcp-pentest