ui-design-system
Fail
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local script
scripts/design_token_generator.pyusing a shell command that interpolates a user-provided hex color string. Without proper sanitization, this pattern allows for command injection attacks where a user provides a malicious string instead of a valid hex code. Evidence is found inSKILL.mdunder Workflow 1 and Workflow 4. - [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes untrusted data (user-provided colors) that is used in a powerful shell command.
- Ingestion points: Brand color input in
SKILL.md. - Boundary markers: Not present; user input is directly concatenated into the command string.
- Capability inventory: The skill utilizes
pythonto run a local script, which has access to the underlying system. - Sanitization: The instructions do not provide any guidance on validating or escaping the hex color input before execution.
Recommendations
- AI detected serious security threats
Audit Metadata