screenshot-to-prototype

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill employs several local Node.js scripts to manage the prototype lifecycle. This includes scripts for image metadata extraction (prepare-reconstruction-source.mjs), manifest building (build-reconstruction-manifest.mjs), and asset auditing (audit-assets.mjs). These scripts operate strictly on local project files and image data.
  • [EXTERNAL_DOWNLOADS]: The skill depends on well-known development packages including vite, tailwindcss, and @tailwindcss/vite to compile styling for the prototypes. These resources are standard in modern web development environments.
  • [DATA_EXFILTRATION]: Analysis of the source code and scripts confirms that image processing and data storage occur within the local project directory (.local/ and src/). No evidence of unauthorized network transmission or access to sensitive system paths (e.g., SSH keys, environment variables) was found.
  • [PROMPT_INJECTION]: The prompt templates provided in references/prompts.md are designed for UI analysis and asset extraction. They do not contain instructions to override safety guardrails or ignore system constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 11:23 AM
Security Audit — agent-trust-hub — screenshot-to-prototype