meti-ai-guideline
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from the Ministry of Economy, Trade and Industry (METI) website and local documents, which could potentially contain malicious instructions if the sources are compromised.\n
- Ingestion points: Content fetched via Playwright/Chrome DevTools from meti.go.jp and files within the references/ directory.\n
- Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded commands within the fetched documents.\n
- Capability inventory: The skill utilizes browser tools for network access, reads/writes local files to update knowledge, and uses the pdftotext utility.\n
- Sanitization: There is no evidence of sanitization or validation of the external content before processing.\n- [EXTERNAL_DOWNLOADS]: The skill fetches guidelines and updates from the official METI website (meti.go.jp), a well-known government service.
Audit Metadata