security-audit-checklist

SUSPICIOUS due to its explicit security-audit/penetration-testing capability for an AI agent, but not malicious. Its footprint is internally consistent with a code-audit checklist, uses standard tools, requests no credentials, and shows no exfiltration or deceptive data flow.