security-engineer

SUSPICIOUS. The skill’s core behavior is mostly aligned with a security-audit purpose and data stays within the local repo, but it equips the agent with offensive security review and penetration-testing capabilities, plus broad autonomous write access for remediation. There is no evident credential harvesting, third-party proxying, or malicious exfiltration path in this fragment, so this is not confirmed malware; however, as an AI agent skill it is high-risk because it enables security scanning/attack-style workflows and substantial autonomous code modification.