frontend-slides

Pass

Audited by Gen Agent Trust Hub on Oct 20, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The generated presentations are designed to fetch assets from trusted and well-known services. Specifically, it uses Google Fonts and Fontshare for typography and Chart.js from a public CDN for data visualization. These downloads are necessary for the intended aesthetic and functionality of the presentations.
  • [COMMAND_EXECUTION]: The skill provides step-by-step instructions for the user to deploy their finished presentations to Vercel and to export them to PDF using Playwright. These commands are intended to be executed by the user in their own environment and are standard workflows for web hosting and browser automation.
  • [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it processes user-provided content (text and image descriptions) to populate slide templates. However, the instructions provide clear boundaries for the agent to prioritize the structural and stylistic rules of the selected design system over embedded data. There is no evidence of malicious data being processed during static analysis.
  • [DYNAMIC_EXECUTION]: The skill generates HTML and JavaScript to handle presentation logic such as slide transitions, viewport scaling, and optional inline editing. The generated code is client-side, focused on UI/UX, and follows standard web development practices without using dangerous execution patterns like eval() on untrusted inputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 20, 2026, 06:45 AM