ui-polish
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external implementation data, 'approved directions', and visual evidence (DesignContext.requiredEvidence) as input.
- Ingestion points: Ingests current implementation code and rendered evidence through the DesignContext.
- Boundary markers: No explicit markers are defined to delimit or sanitize instructions that might be embedded in the implementation source code or design documentation.
- Capability inventory: The skill instructs the agent to use the host's existing edit, run, browser, and screenshot capabilities.
- Sanitization: No sanitization or validation of the input content is specified before the agent acts on the implementation.
Audit Metadata