design-lab
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests design tokens and patterns from project files such as
tailwind.config.js,package.json, and existing UI components. This creates a surface where malicious instructions embedded in project source code could influence the agent's behavior. \n - Ingestion points: Reads project configuration and component source files (SKILL.md Phase 0).\n
- Boundary markers: No explicit delimiters or warnings to ignore embedded instructions are defined for the ingested project data.\n
- Capability inventory: The skill has the capability to write new files, modify routes, and delete directories (SKILL.md Phase 3 and 8).\n
- Sanitization: No specific sanitization or validation of the inferred style data is performed before it is used in code generation.\n- [DYNAMIC_EXECUTION]: The skill generates multiple React components and temporary framework routes at runtime. This involves programmatically assembling UI code based on a generated Design Brief and user feedback.\n- [COMMAND_EXECUTION]: The skill probes the project environment to detect package managers and framework types by checking for the existence of specific lock files and configuration files. It also provides instructions for users to run local development servers.
Audit Metadata