vc-curated-match
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted product descriptions and URLs via CLI arguments in
scripts/run.pyto generate matching reports. \n - Ingestion points: The
--descriptionand--urlparameters in the orchestrator script receive data from external user/agent context. \n - Boundary markers: No specific boundary markers are used in the generated Markdown report to isolate user-supplied content. \n
- Capability inventory: The skill possesses file-writing capabilities to the local filesystem but does not perform network operations. \n
- Sanitization: The
scripts/generate_report.pyscript escapes square brackets in user strings to prevent Markdown link hijacking, but it does not filter for embedded prompt instructions. \n- [COMMAND_EXECUTION]: The skill provides a CLI interface that accepts a user-defined--outputpath, allowing it to write the results of the VC matching process to an arbitrary file location within the agent's permitted environment.
Audit Metadata