icp-deep-scanner

SUSPICIOUS: the stated purpose is coherent, and there is no direct evidence of malware, credential theft, or hostile routing to attacker infrastructure. However, the capability footprint is broad for a read-mostly research skill: it can read many sensitive systems, fetch large amounts of untrusted external content, dispatch sub-agents, and write outputs, creating medium risk from prompt injection, overscoped access, and transitive trust in related skills.