The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to interact with and process data from external, untrusted web applications which could contain malicious payloads designed to influence agent behavior.
Ingestion points: The skill fetches HTTP headers, server responses, JavaScript bundles, and API data from user-specified target domains using the WebFetch tool as described in Phase 1 through Phase 9 of the methodology.
Boundary markers: The instructions lack specific boundary markers or 'ignore' commands to prevent the agent from accidentally executing instructions found within the HTML, headers, or error messages of the target application.
Capability inventory: The skill employs Bash (executing tools like curl, testssl.sh, and sslyze), Write (to generate reports), and WebFetch (to retrieve external content).
Sanitization: There are no sanitization or validation routines provided for the external data retrieved before it is processed or written to the report file.
[EXTERNAL_DOWNLOADS]: The skill provides links to external resources for security methodology.
Finding: The # References section in SKILL.md includes a link to book.hacktricks.xyz. While this is a common security research site, it is flagged by some reputation scanners because it contains proof-of-concept payloads for security vulnerabilities.

web-pentest