design-anti-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to search existing project files for color palette definitions to maintain design consistency, which creates a potential surface for indirect prompt injection if project assets contain malicious instructions.
- Ingestion points: Project files are scanned for existing color palettes as described in the 'Color Selection' sections of SKILL.md and references/uncodixfy-rules.md.
- Boundary markers: The skill does not provide specific delimiters or instructions to ignore potential natural language commands within the color data or project files.
- Capability inventory: The skill is authorized to generate complex frontend code (HTML, CSS, React, etc.) across multiple files, providing a clear path for injected instructions to influence generated code.
- Sanitization: There is no requirement or instruction to validate the ingested color data against a schema or to sanitize it before interpolation into the generation process.
Audit Metadata