prototype

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill generates and writes executable code (HTML, JavaScript, and CSS) to the local file system to provide a live prototyping experience. It mitigates potential risks by using isolated directories and providing strict instructions to delete these files once a variant is selected.
  • [INDIRECT_PROMPT_INJECTION]: The skill operates by ingesting user-provided descriptions to define the UI variants. * Ingestion points: User descriptions in Phase 1 and project-specific token/context files in Phase 2. * Boundary markers: None explicitly specified for the user description, but logic is constrained by a strict design engineering persona and phase-based workflow. * Capability inventory: File-write capabilities are used to create isolated prototype routes or standalone files; no network access is requested. * Sanitization: The skill relies on the AI model's generation logic without explicit input filtering, but isolates the output from production code.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 10:24 PM
Security Audit — agent-trust-hub — prototype