deep-research-offensive

SUSPICIOUS. The skill’s capabilities mostly match its stated purpose as an offensive research workflow, but that purpose itself is high-risk: it guides an AI agent to gather exploit intelligence, PoCs, recon sources, and attack-chain material. There is no confirmed malware, no secret exfiltration, and no pre-execution payloads, but the workflow expands network reach to multiple third-party proxies and optionally forwards real X credentials to twikit. Overall this is a coherent but security-sensitive offensive skill with medium-high operational risk rather than confirmed malicious content.