starlight-design-review
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides instructions for design auditing and does not contain malicious code or unsafe command patterns.
- [SAFE]: Tool usage is restricted to host-approved environments, explicitly prohibiting arbitrary binary execution.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes live previews of user-facing surfaces. 1. Ingestion points: Live previews or runnable surfaces provided as input (SKILL.md). 2. Boundary markers: No explicit boundary markers or 'ignore' instructions are defined for the input data. 3. Capability inventory: Uses host-approved browser or design connectors; no network exfiltration or command execution capabilities detected. 4. Sanitization: No specific sanitization or filtering of the ingested visual/semantic content is described.
Audit Metadata