The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides numerous command-line templates for executing reconnaissance tools, including curl, ffuf, gobuster, bfac, katana, hakrawler, gau, and x8 to perform directory discovery, parameter fuzzing, and endpoint crawling across several phases of methodology described in SKILL.md and references/web-discovery-methods.md.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze data from external, untrusted web targets, creating a surface for indirect injection attacks.
Ingestion points: The agent is instructed to process HTTP headers, status codes, HTML/JS comments, and JavaScript bundles from remote targets via tools like httpx, nuclei, katana, and hakrawler (detailed in Phase 3, Phase 4, and the reference guide).
Boundary markers: Absent. There are no instructions for the agent to use delimiters or specifically ignore instructions embedded in the target's response data.
Capability inventory: The skill utilizes network tools to identify technology stacks and vulnerabilities, and then instructs the agent to map these findings to specific attack strategies, potentially executing further scans based on that data.
Sanitization: No evidence of sanitization, escaping, or filtering of content fetched from external targets before it is presented to the agent for analysis.

recon-full