yara-authoring

SUSPICIOUS due to dual-use security tooling for an AI agent, but not malicious. The skill is internally consistent with YARA rule authoring, uses no credentials, has no exfiltration path, and shows no untrusted installer or third-party proxy behavior.