web-security-testing

SUSPICIOUS/HIGH-RISK workflow. The content is coherent with a pentesting purpose, but that purpose itself grants an AI agent offensive security capabilities and delegates to unspecified external skills without provenance or authorization guardrails. No direct malware or exfiltration is shown in this file, but the transitive trust and autonomous attack guidance make it unsafe to classify as benign.