payloads-file-upload

This is an exploit-focused security skill, not a passive reference: it equips an AI agent to test upload endpoints with payloads that can achieve command execution, file disclosure, and outbound connections. No evidence shows credential theft, concealed behavior, or automatic execution, so malicious intent is not confirmed; nevertheless, the offensive capability and lack of enforced per-action authorization create high security risk.