claims-and-proof-check
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze user-supplied marketing copy for regulatory claims. While this represents a surface for untrusted data to enter the agent's context, the analysis is performed by a bundled local script using regular expressions rather than relying solely on the LLM to identify patterns, which mitigates standard injection risks.
- [COMMAND_EXECUTION]: The skill uses a local Node.js script (
resources/claims-check.js) to perform text processing. The script is invoked via a standard CLI command provided in the instructions. Analysis of the script confirms it contains only local logic for string manipulation and regex matching, with no network access, sensitive file operations, or dynamic code execution capabilities. - [EXTERNAL_DOWNLOADS]: The skill references multiple external URLs for regulatory guidance (e.g., ftc.gov, hhs.gov, ed.gov) and vendor-owned resources (aajconsult.com). All referenced URLs are presented in plain text and target official government portals or the publisher's established domain, posing no risk to the execution environment.
Audit Metadata