scrape
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted job descriptions from the web and external job portals, which could contain instructions to manipulate the agent's behavior during fit assessment.\n
- Ingestion points: Job details fetched via WebFetch and portal CLIs in Step 2.\n
- Boundary markers: The skill does not define specific delimiters or instructions to isolate external job content from its own logic.\n
- Capability inventory: The skill can write to local JSON files, execute specific Python scripts, and perform web searches.\n
- Sanitization: No explicit sanitization or escaping of external content is mentioned before it is processed by the agent.\n- [COMMAND_EXECUTION]: The skill discovers and initiates execution of command-line tools by reading and parsing
SKILL.mdfiles from other directories in the.agents/skills/path.\n - Evidence: Step 1b explicitly instructs the agent to read
SKILL.mdfiles to find the correctbun runinvocation and flags.\n- [DYNAMIC_EXECUTION]: The search functionality relies on dynamically constructing and executing commands based on the content of external skill metadata files encountered at runtime.
Audit Metadata