ctf-pwn

SUSPICIOUS: the skill is internally coherent, but its purpose is to give an AI agent offensive binary-exploitation capability. There is little evidence of malware or credential theft, yet the exploit-development functionality itself makes this a high-security-risk skill.