design-theme
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses standard shell commands such as cat, grep, and rg to inspect local project configuration files (e.g., tailwind.config.*, globals.css). These actions are performed to discover existing design tokens and ensure the new theme integrates correctly without breaking existing styles.
- [INDIRECT_PROMPT_INJECTION]: The skill processes project files which constitutes an indirect prompt injection surface. 1. Ingestion: Local project files are read via shell commands in SKILL.md. 2. Boundary markers: No explicit delimiters are used for the ingested content. 3. Capability: The agent is authorized to write stylistic changes (themes) to artifacts. 4. Sanitization: No specific filtering of the ingested file content is performed. This surface is considered a standard risk for development tools and is functionally safe given the specialized task of style extraction.
Audit Metadata