phone-ui-demos

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill workflow includes a quality check instruction (Step 2 in SKILL.md) that uses a local command line: node ../video-shot-demos/scripts/shot.js 页面.html 6000 _t6s.png. This pattern assumes the presence of a sibling directory containing the specified script for automated screenshot generation.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied data to generate demonstration content, creating an attack surface common to content generation tools.\n
  • Ingestion points: User-provided product functional lists, knowledge outlines, or scripts (SKILL.md, Step 0).\n
  • Boundary markers: Absent. The generated code interpolates user content directly into HTML elements (e.g., chat bubbles, notifications) without structural delimiters or isolation markers.\n
  • Capability inventory: The generated HTML assets use Web Audio API and CSS 3D transitions. The overall workflow suggests the use of a local shell tool for automated quality assurance.\n
  • Sanitization: Absent. The skill instructions do not mandate escaping or sanitizing user-provided text before it is inserted into the HTML and JavaScript templates.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 11:46 AM
Security Audit — agent-trust-hub — phone-ui-demos