solidity-audit

SUSPICIOUS. The skill is coherent with its stated purpose, and its main tools map to real official ecosystems, so there is no clear malware or credential-harvesting signal. However, it is a high-risk security/exploit skill for an AI agent: it reads untrusted code, executes analyzers and tests, spawns sub-agents, and can generate exploit PoCs, which materially increases operational risk even though data exfiltration and credential scope appear low.