authoring-motion-diagrams

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requires running project-local scripts to validate and generate diagram assets.\n
  • Evidence: The authoring flow in SKILL.md uses node scripts/animations/render.js --check, node scripts/animations/preview.js, and npm test to verify scenes.\n
  • Evidence: Documentation in mechanic.md describes using headless Chrome for frame capture and FFmpeg for video encoding.\n\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests repository-local configuration data to guide its actions, which represents an attack surface for indirect injection.\n
  • Ingestion points: Step 1 of the flow in SKILL.md instructs the agent to read plugin categories from .claude-plugin/marketplace.json and color tokens from scripts/animations/palette.js.\n
  • Boundary markers: No explicit boundary markers or delimiters are defined for these data sources in the instructions.\n
  • Capability inventory: The agent has capabilities for file system writes (SVG scenes) and execution of local Node.js scripts.\n
  • Sanitization: The render.js --check tool validates design constraints (like grid alignment and contrast) but does not provide security-focused sanitization of the input data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 08:49 PM
Security Audit — agent-trust-hub — authoring-motion-diagrams