web-attack-methods

High-risk offensive-security skill. Its capabilities are internally consistent with a red-team purpose, but that purpose itself gives an AI agent concrete exploitation, bypass, and exfiltration instructions against external systems. No strong evidence of hidden malware or deceptive credential harvesting from the user, but the skill materially increases the chance of harmful autonomous abuse.