pentest-recon-surface-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to perform reconnaissance, which involves ingesting data from external, potentially untrusted targets during asset discovery.\n
- Ingestion points: Target URLs, wordlists, and HTTP responses from services being scanned (SKILL.md).\n
- Boundary markers: The instructions lack specific delimiters or warnings to ignore instructions that might be embedded in target responses.\n
- Capability inventory: The skill provides templates for executing shell commands using the
ffufutility (SKILL.md).\n - Sanitization: No sanitization or validation logic is specified for the data gathered from external targets before it is used to influence subsequent testing phases.\n- [COMMAND_EXECUTION]: The skill provides a guide for executing the
ffuftool via the command line to perform virtual host fuzzing. It uses a variable$TARGETwhich is interpolated into the shell command.
Audit Metadata