The Agent Skills Directory

[SAFE]: The skill is a specialized utility for web security auditing. Its primary function is to detect auto-generated index pages (autoindex) on web servers by comparing HTTP response bodies against a known database of fingerprints for servers like Apache, nginx, and Caddy. All network activity is directed toward user-provided and authorized targets.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external web servers, which creates a potential surface for indirect prompt injection.
Ingestion points: The Python script scripts/probe_directory_listing.py ingests untrusted data from the resp.text field after performing GET requests on remote targets.
Boundary markers: The findings are reported back to the agent without explicit delimiter-based isolation for the external content.
Capability inventory: The skill utilizes Bash(python3:*) and Bash(curl:*) to perform its network operations and analysis.
Sanitization: The script performs validation by checking the first 4KB of the response body against static regex signatures (AUTOINDEX_PATTERNS) before identifying a finding, which helps mitigate accidental processing of malicious content.

detecting-directory-listing