translator
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No executable code, shell commands, or network access detected. The skill is purely instructional for LLM behavior and operates within the context of text transformation.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text for translation. 1. Ingestion point: User input for translation processed by instructions in SKILL.md. 2. Boundary markers: The prompt instructs the agent to use Markdown headers and code blocks for output isolation. 3. Capability inventory: No subprocess, file system, or network tools are enabled or used. 4. Sanitization: No explicit content filtering or escaping of user input is specified. The risk is minimal as the agent has no dangerous capabilities to be exploited.
Audit Metadata