translate
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements a structured orchestration pattern for a translation suite. It relies on a predefined list of local sub-skills with hardcoded relative paths, minimizing the risk of unauthorized file access or path traversal.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted user input for translation purposes, which represents a standard injection surface. However, the presence of explicit instructions forbidding external data storage and unauthorized actions significantly mitigates potential risks.
- Ingestion points: The skill processes the 'original message' and 'conversation context' as core data for its translation operations (SKILL.md).
- Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded instructions in the user text within this coordination layer.
- Capability inventory: The skill is limited to reading instructions from a fixed set of local files and passing data between them; it specifically instructs the agent not to write data to external storage.
- Sanitization: There is no evidence of input validation or sanitization for the user-provided text before it is processed by the translation sub-skills.
Audit Metadata