source-html-to-app-ui
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a custom utility script, playwright-lifecycle.mjs, to execute shell commands (bash -lc) for setting up environments and starting local servers.
- [DYNAMIC_EXECUTION]: The workflow involves the creation and execution of custom Playwright scripts for UI verification, managed by a provided lifecycle helper.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external HTML and JavaScript, representing an injection surface. 1. Ingestion points: Source HTML files and design inputs are copied into the task-workflow/ directory. 2. Boundary markers: A mandatory process gate (First-action lock) requires orientation images before any code reading occurs. 3. Capability inventory: Capabilities include shell command execution via playwright-lifecycle.mjs, filesystem access, and browser interaction. 4. Sanitization: Verification is performed through process-level visual review gates rather than code sanitization.
- [SAFE]: The skill demonstrates secure practices, such as preventing the override of critical environment variables and restricting repository access until a specific checkpoint is passed.
Audit Metadata