ux-writing
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run a shell command (ls) to check for the presence of a 'tokens' directory. This is used as an environment check to verify the presence of a design system kit.
- [INDIRECT_PROMPT_INJECTION]: The skill processes content from local project documentation files to inform its copy suggestions.
- Ingestion points: Reads files located at project-relative paths for voice, tone, and accessibility guidelines.
- Boundary markers: None explicitly defined in the instructions for the read files.
- Capability inventory: The skill performs directory checks and file reads; it does not demonstrate file-write or network capabilities.
- Sanitization: No specific sanitization of the content from ingested documentation is mentioned.
Audit Metadata