testing-for-business-logic-vulnerabilities

SUSPICIOUS: the skill is internally coherent and uses official/normal tools, but its actual purpose is to equip an AI agent with offensive penetration-testing techniques that can manipulate transactions, exploit race conditions, and attempt privilege escalation. There is no clear credential-harvesting or malware behavior, yet the capability itself is high risk for agent deployment and should be treated as a dangerous security-testing skill rather than benign workflow guidance.