The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local shell commands including curl, jq, mkdir, grep, sed, and wc to interact with target environments and process scan results. These operations are standard for security reconnaissance and do not include dangerous patterns like piping remote content directly to a shell.
[EXTERNAL_DOWNLOADS]: Network operations are performed via curl to fetch configuration data and assets from user-defined target URLs. All requests are directed to the target or specified test origins (attacker.example) and are used for protocol classification and version detection.
[SAFE]: The skill incorporates extensive ethical guidelines and prerequisites, including mandatory authorization, boundary definitions, and negative control testing. It correctly identifies that reconnaissance signals are not findings in themselves and cautions against state-changing tests without explicit approval.

hunt-wordpress