The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a bash script (recon.sh) that automates the execution of multiple external security tools including subfinder, amass, dnsx, httpx, naabu, whatweb, gowitness, and h8mail. It also includes network routing commands such as sudo ip route add used during pivoting operations to access internal networks.
[EXTERNAL_DOWNLOADS]: The technical deliverables reference and require several third-party offensive security utilities such as chisel, ligolo-ng, and h8mail for their operation.
[PROMPT_INJECTION]: The skill adopts a specialized offensive persona; however, it incorporates explicit "Critical Rules" regarding engagement authorization, scope management, and ethical standards to ensure actions are performed legally and safely.
[DATA_EXFILTRATION]: The skill methodology includes techniques for checking credential leaks via h8mail and documents data exfiltration patterns as part of its core offensive security mission.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it is designed to ingest and process untrusted data from external network targets.
Ingestion points: Data enters the agent's context through network responses captured by the requests library in the Python script and probing tools in the bash script.
Boundary markers: The instructions lack specific markers or delimiters to differentiate between target data and agent instructions when analyzing external responses.
Capability inventory: The skill environment allows for shell command execution, file system interaction, and network connectivity.
Sanitization: The provided scripts do not include mechanisms for filtering or sanitizing external content before it is processed by the agent.

agency-penetration-tester