smart-contract-auditing

SUSPICIOUS. The skill is internally coherent as a smart-contract auditing guide: its capabilities, examples, and tooling match the stated purpose, and there is no evidence of credential harvesting, exfiltration, hidden execution, or untrusted install chains inside the skill itself. However, it materially enables offensive security behavior by teaching exploit development and attack simulation, so it should be treated as high-risk security tooling for an AI agent even though it does not appear malicious.