swiss-user-research-video-template
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions and provided templates in assets/template.html and example.html are entirely self-contained. No external dependencies, libraries, or scripts are loaded from the network during execution or rendering.\n- [SAFE]: The skill workflow and internal checklist in references/checklist.md explicitly prohibit the use of sandbox-hostile browser APIs such as localStorage, alert, and window.open, maintaining a restricted and safe execution environment for the generated artifacts.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing user-supplied research data into HTML artifacts. The risk is minimized by structural constraints and explicit restrictions on script capabilities. 1. Ingestion points: User research data mapped to template placeholders. 2. Boundary markers: Workflow requires a specific three-slide structure. 3. Capability inventory: Primary capability is file_write for the resulting HTML artifact. 4. Sanitization: The checklist provides explicit filtering by forbidding dangerous browser APIs.
Audit Metadata