master-xuanzang
Pass
Audited by Gen Agent Trust Hub on May 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill establishes a high-fidelity persona for Buddhist scholarly dialogue, employing rigorous formatting and citation rules to ensure the accuracy of doctrinal claims against historical canons.\n- [PROMPT_INJECTION]: The skill utilizes internal behavioral constraints, referred to as 'HARD-GATE' and 'Iron Rules', to prevent the agent from providing teaching without citations or making sectarian judgments. These instructions are designed for persona alignment and accuracy rather than attempting to bypass the underlying model's safety filters.\n- [COMMAND_EXECUTION]: The skill documentation references optional auxiliary scripts (
scripts/cite.pyandscripts/query.py) for retrieving canonical citations. These are presented as internal utility tools for the agent's research process and do not involve untrusted remote code execution or dangerous shell commands.
Audit Metadata