skills/editframe/skills/composition/Gen Agent Trust Hub

composition

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill defines mechanisms to ingest external data for video customization through window.EF_RENDER_DATA and remote caption files via the captions-src attribute. This establishes a surface where untrusted data could influence agent actions or rendering output. 1. Ingestion points: window.EF_RENDER_DATA, captions-src attribute, captions-script ID. 2. Boundary markers: Not specified. 3. Capability inventory: JavaScript execution via initializer and addFrameTask on temporal elements. 4. Sanitization: Not explicitly documented; implementers are responsible for input validation.
  • [DYNAMIC_EXECUTION]: The composition model permits custom JavaScript to be executed on a per-frame basis using initializer and addFrameTask hooks to drive animations and computed state within the rendering pipeline.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes several official Node.js packages under the @editframe scope and references configuration settings for the vendor's API host and media signing services.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 11:15 AM
Security Audit — agent-trust-hub — composition