pentest-web-enumeration

SUSPICIOUS/HIGH-RISK but not confirmed malware. The skill is internally consistent with its stated pentest purpose and mostly uses official tooling, but it gives an AI agent offensive web-enumeration and early vulnerability-scanning capability, reads local credential material, and processes untrusted web content while executing tools and writing artifacts. Main risk is agent-enabled security testing against live targets, not hidden exfiltration or deceptive supply-chain behavior.