motion-graphics

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPRIVILEGE_ESCALATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill fetches and processes untrusted data from the internet for categories such as news, tweet, and webpage. This creates a surface for indirect prompt injection where malicious instructions embedded in external content could attempt to influence the agent's planning or building phases.
  • Ingestion points: External content is fetched in phases/source/guide.md and processed by the Director subagent in Step 3 of the flow.
  • Boundary markers: There are no explicit instructions or delimiters provided to the subagents to ignore or isolate instructions found within the fetched source material.
  • Capability inventory: The skill can execute shell commands (via ffmpeg and hyperframes CLI) and generate HTML/JavaScript compositions.
  • Sanitization: The Director subagent extracts specific fields (e.g., headline, keyword), providing a level of structural filtering, but no formal sanitization or safety-specific constraints for the input text are defined.
  • [COMMAND_EXECUTION]: The skill makes extensive use of the hyperframes CLI tool via npx for initialization, rendering, linting, and inspection tasks. Additionally, it executes system tools like ffmpeg, ffprobe, and base64 using Node.js child_process modules in bake-basemap.mjs and locate.mjs.
  • [EXTERNAL_DOWNLOADS]: The skill references and downloads several resources from well-known and trusted services, which is generally considered safe:
  • Map-related libraries (maplibre-gl, topojson-client) and datasets (world-atlas) are fetched from cdn.jsdelivr.net within bake-basemap.mjs.
  • The GSAP animation library is loaded from cdn.jsdelivr.net in the circle-highlight sample template.
  • The npx hyperframes init command performs updates by checking against a remote repository on GitHub.
  • [PRIVILEGE_ESCALATION]: The categories/maps/bake-basemap.mjs script launches a headless Puppeteer browser using the --no-sandbox flag. While frequently required for execution in CI or containerized environments, this disables a primary security boundary when rendering content from external tile servers or CDN-hosted libraries.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 12:01 AM