motion-canvas-agent
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill extracts and serializes text content from the Motion Canvas scene graph to provide the agent with project context, creating a surface for indirect prompt injection.
- Ingestion points: The
serializeNodefunction inassets/agent-client.tsreads text directly from animation nodes usingnode.text(). - Boundary markers: The skill provides no explicit boundary markers or instructions to the agent to disregard embedded directives within the scene graph JSON.
- Capability inventory: The skill possesses file system write capabilities (
fs.writeFileSyncinassets/agent-plugin.ts) and full control over the animation player/renderer. - Sanitization: The skill performs basic sanitization by stripping control characters and truncating text to 100 characters, but this does not prevent instructions written in plain text from being processed by the agent.
Audit Metadata