conducting-internal-network-penetration-test

SUSPICIOUS: the skill is internally consistent as an internal pentest playbook, but it equips an AI agent with high-risk offensive capabilities including credential harvesting, remote execution, privilege escalation, and domain compromise. No obvious third-party credential-harvesting endpoint is shown, but the skill is dangerous by design and inappropriate for broad autonomous use.