The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill operates on external, untrusted data sources (such as DNS records, web server banners, and certificate logs), creating an attack surface where malicious target content could attempt to influence the agent's behavior.
Ingestion points: Data enters the agent's context through OSINT gathering and active service enumeration (SKILL.md).
Boundary markers: Present. The skill includes mandatory pre-engagement requirements and non-negotiable rules that emphasize written authorization and defined scope (SKILL.md).
Capability inventory: The skill utilizes command-line tools for network scanning, service discovery, and vulnerability exploitation (SKILL.md).
Sanitization: Absent. The instructions do not define specific mechanisms for escaping or filtering external content before it is processed or included in reports.
[COMMAND_EXECUTION]: The skill provides specific command-line templates for standard security auditing and penetration testing tools.
Evidence: Detailed command strings for utilities such as nmap, gobuster, nikto, and sqlmap are provided for the agent to execute during authorized testing phases (SKILL.md).

penetration-tester