master-daoxuan
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill uses
curlto retrieve Buddhist textual data fromfojin.appwhen local documentation is incomplete. This is a functional requirement for its purpose as a scholarly teaching assistant and is limited to HTTP GET requests for data retrieval. - [INDIRECT_PROMPT_INJECTION]: The skill recognizes the risk of ingesting third-party content from external APIs. It mitigates this by requiring the agent to wrap retrieved content in
<<<FOJIN_DATA>>>boundary markers and explicitly instructs the agent to treat the text as data only, ignoring any embedded instructions or overrides. - [SAFE]: Extensive safety 'Hard Gates' are implemented to ensure the AI does not attempt to diagnose mental health issues, perform religious ceremonies (e.g., ordinations), or certify personal spiritual achievements, instead referring users to qualified human professionals and clergy.
Audit Metadata