hunt-cloud-misconfig

SUSPICIOUS. The skill is internally coherent as an offensive cloud-misconfiguration hunting playbook, but its actual footprint is high-risk for an AI agent: it teaches credential extraction, unauthorized cloud enumeration, write validation on third-party resources, covert exfiltration through AWS RUM, and chaining into other attack skills. Install provenance is mostly official and there is no strong malware signal, but the autonomy and exploit guidance make the skill dangerous and disproportionate for routine agent use.