bazi-ziwei
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied birth information and fortune-telling chart text, which is then passed to the LLM for analysis. This creates a surface for indirect prompt injection if the user-provided text contains malicious instructions aimed at the analysis phase.
- Ingestion Points: User birth details and user-provided chart data (Step 0 in SKILL.md).
- Capability Inventory: The skill performs file writing and template rendering via local TypeScript scripts executed by the agent.
- Boundary Markers: None explicitly defined in the prompts to separate user data from analysis instructions.
- Sanitization: The skill relies on structured input for birth dates but does not explicitly sanitize free-text chart data provided by the user.
Audit Metadata