attack-surface-recon
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes several command-line tools for security testing, including amass, subfinder, dnsx, httpx, naabu, nmap, katana, gau, ffuf, and nuclei. These are legitimate tools used for the skill's primary function of infrastructure reconnaissance.
- [INDIRECT_PROMPT_INJECTION]: The skill consumes data from external sources and target web resources, creating a potential surface for indirect prompt injection.
- Ingestion points: Commands querying crt.sh, web.archive.org, and performing content discovery on target subdomains (SKILL.md).
- Boundary markers: None specific; however, data flows through standard shell pipes which keeps the processing distinct from the agent's control flow.
- Capability inventory: The skill possesses capabilities for network interaction, file writing, and arbitrary command execution as defined in the reconnaissance scripts (SKILL.md).
- Sanitization: Output is filtered using text processing tools like grep, sort, and head to isolate specific identifiers like domain names.
- [REMOTE_CODE_EXECUTION]: A command in SKILL.md uses curl to fetch JSON data and pipes it to python3. Analysis confirms this is a data-processing step where a local Python script parses the received JSON, rather than the execution of untrusted remote code.
Audit Metadata