The Agent Skills Directory

[SAFE]: The skill acts as a security gate to prevent unauthorized actions. It defines a mandatory 11-point checklist for Rules of Engagement (ROE) and mapping to international security standards (NIST, OWASP, ISO). No malicious patterns were detected.
[COMMAND_EXECUTION]: The skill explicitly forbids the use of execution tools such as Bash via its frontmatter configuration.
[DATA_EXFILTRATION]: The skill forbids network operations (WebFetch) and contains explicit rules to ignore sensitive files like .env, secrets, and private keys. It also implements redaction policies for credentials and PII in its output.
[REMOTE_CODE_EXECUTION]: There are no patterns suggesting the download or execution of remote code.
[INDIRECT_PROMPT_INJECTION]: While the skill processes untrusted user input regarding penetration testing scopes, it lacks the capabilities (write access, command execution, or network access) required to exploit such injections. It functions as a read-only planning assistant.
[PRIVILEGE_ESCALATION]: No attempts to escalate privileges or bypass platform constraints were identified.

pentest-scope-and-roe