The Agent Skills Directory

[SAFE]: The skill is strictly read-only and explicitly forbids the agent from performing write operations, executing shell commands, or making network requests.
[SAFE]: It enforces a 'Double-Gate' authorization model, requiring a verified Rules of Engagement document and explicit human consent in the current session before any scanning activities are planned.
[SAFE]: The instructions prohibit scanning production environments and third-party authentication services, focusing exclusively on authorized testing targets.
[SAFE]: Authenticated scanning is restricted to dedicated test accounts, with a strict prohibition against using real administrative or user credentials.
[SAFE]: All findings are required to undergo automated redaction of session tokens and personal information to prevent sensitive data exposure in logs and reports.

dast-authenticated