design-system
Pass
Audited by Gen Agent Trust Hub on Jul 18, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted external data such as UI screenshots and component source code to extract design rules, which represents an attack surface for indirect prompt injection. \n
- Ingestion points: External evidence is ingested during discovery in SKILL.md (Workflow Step 3) and described in workflow.md (Evidence boundary). \n
- Boundary markers: The skill enforces mandatory 'use' and 'ignore' lists for external references in SKILL.md (Workflow Step 5) and includes hard rules against inventing functionality or features from visual sources. \n
- Capability inventory: The skill is restricted to repository inspection and local file validation using the included scripts/validate_ui_artifacts.py. Actions like browser interaction are delegated to separate tools. \n
- Sanitization: The provided validation script uses yaml.safe_load() to ensure secure processing of YAML-formatted artifacts. \n- [COMMAND_EXECUTION]: The skill utilizes repository-bound commands such as 'git status' and a local Python validation script. These operations are limited to the local environment and are strictly necessary for the skill's design system management purpose.
Audit Metadata