top-web-vulnerabilities

SUSPICIOUS: the skill is internally coherent as a web-security reference and shows no credential harvesting, hidden data flow, or suspicious installer behavior. However, it equips an AI agent with actionable penetration-testing guidance, including verification techniques and troubleshooting for bypassing defenses, which makes it high-risk as an agent skill despite low malware evidence.