translate-article
Pass
Audited by Gen Agent Trust Hub on Feb 21, 2026
Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
- Indirect Prompt Injection (LOW): The skill processes untrusted content from documentation files without using boundary markers, which could allow malicious text within a source file to influence the agent's behavior.
- Ingestion points: MDX files within src/content/docs/.
- Boundary markers: Absent.
- Capability inventory: File system write access for localization.
- Sanitization: Absent.
- No Code (SAFE): This skill contains no executable scripts or binary files, significantly reducing the technical attack surface for malware or persistence.
Audit Metadata