uit-actions-web
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions define a surface for indirect prompt injection by detailing UI components that ingest untrusted data.
- Ingestion points: The InputTextView component and customizable button labels/subtitles (referenced in SKILL.md and references/action-components.md) serve as ingestion points where external data enters the agent context.
- Boundary markers: The provided instructions do not include requirements for using delimiters or boundary markers to distinguish user-provided content from agent instructions.
- Capability inventory: The skill enables actions that trigger product capabilities, application state changes, and navigation, which could be misdirected by instructions embedded in UI data.
- Sanitization: No explicit requirements or patterns for sanitizing or escaping external content before display are mentioned.
- [SAFE]: The skill includes strong security guidelines for sensitive data handling, explicitly instructing that secret values like PINs and access codes must be masked or omitted from rendered and accessible content even when labeled as hidden.
Audit Metadata