The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes shell commands including curl, grep, shasum, and cat to identify technology stacks and perform baseline 404 checks on target hosts. These operations are restricted to reconnaissance tasks consistent with the skill's purpose as a penetration testing tool. Evidence is found in the bash scripts within Step 0 and Step 1 of SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes untrusted data from external web servers to determine subsequent actions.
Ingestion points: HTTP response headers and bodies fetched from targets via curl in Step 1 of SKILL.md.
Boundary markers: The skill does not implement specific delimiters for the processed content, although it filters data through specific regex patterns in Step 1.
Capability inventory: File reads (cat), network requests (curl), and shell utilities (grep, shasum) are used throughout SKILL.md.
Sanitization: Content is filtered using grep -aoE to extract specific technology markers before being used to influence skill loading logic.

hunt-dispatch