ui-parity
Pass
Audited by Gen Agent Trust Hub on Nov 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: No malicious patterns or security risks were identified in the skill. The instructions emphasize strict adherence to host safety rules, data ownership, and isolation.- [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection as it ingests and processes untrusted content from external websites (implementation and reference UIs) via browser automation tools.
- Ingestion points: Browser captures and DOM facts are collected from external URLs specified in the orchestration brief.
- Boundary markers: The skill includes instructions to distinguish observed facts from intent and to escape content in generated reports.
- Capability inventory: The skill uses
playwright-clifor browser automation and is authorized to perform file-system writes for implementation repairs. - Sanitization: Report generation instructions in
references/evidence.mdexplicitly mandate escaping labels and captions to prevent script injection in the final report artifact.
Audit Metadata