meridian-doc-consultant
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- Retrieval of External Documentation: The skill is configured to fetch technical guides from the official documentation domain. This process is used to provide the agent with up-to-date context on the Meridian framework. As the source is a trusted official documentation site, this is a standard and safe operational pattern.
- Utility Tool Usage: The instructions describe using
curlto download plain-text versions of documentation. This tool is employed solely for data ingestion and does not involve executing the retrieved content or performing unauthorized system modifications. - Potential for Indirect Prompt Injection: Like any skill that ingests external data, there is a technical surface for indirect prompt injection if the source content were to be manipulated. The skill manages this by focusing on specific trusted documentation pages and requiring the agent to ground its responses strictly in the retrieved text.
Audit Metadata