mcp-server-patterns
Audited by Socket on Sep 19, 2026
2 alerts found:
Securityx2SUSPICIOUS: The skill’s purpose and capabilities are internally coherent, and its install/data-flow guidance is mostly proportionate and official. However, it meaningfully equips an AI agent with offensive security and security-testing workflows, including wrapping tools like nmap/sqlmap into callable APIs and discussing stealth/evasion patterns, which makes the overall skill high-risk even without clear credential theft or malicious exfiltration.
No clear malware or intentional supply-chain sabotage is present. The fragment is security-tool integration guidance with significant potential security risk if implemented as shown, especially shell command injection through the nmap command builder and unrestricted offensive-tool inputs. Use execFile with separate arguments, strict allowlists, authorization, target scope controls, path restrictions, timeouts, and isolated execution before deploying.