better-interface
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes repository files such as documentation and source code to perform its review, creating a surface for indirect instructions. 1. Ingestion points: Project documentation files (e.g., CONTRIBUTING.md, CLAUDE.md) and source code files. 2. Boundary markers: The skill does not define specific prompt delimiters but mandates a read-only stance for all reviews to prevent unauthorized changes. 3. Capability inventory: File reading, optional file writing (upon user request), and execution of local project commands. 4. Sanitization: None specified for ingested file content.
- [COMMAND_EXECUTION]: The instructions permit the agent to identify and run preview or test commands already established within the project's own framework to verify runtime UI behavior. This capability is restricted to tools and scripts already present in the user's workspace.
Audit Metadata