exploit-chain

The skill is internally consistent with its stated pentest purpose and does not install tools, request credentials, or route data to third parties. However, it equips an AI agent with offensive security methodology for chaining vulnerabilities into higher-impact attack paths, which makes it high-risk as a capability even without direct exploit code. Overall: suspicious/high-risk offensive security guidance, but not malware.