rendering
Pass
Audited by Gen Agent Trust Hub on Jul 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides legitimate instructions for Blender rendering workflows, including engine selection (Cycles/Eevee), camera setup, and render pass management.
- [SAFE]: The Model Context Protocol (MCP) integration is used for standard development tasks such as configuring scene settings, capturing viewport screenshots (
get_viewport_screenshot), and saving project files (save_mainfile). - [SAFE]: The workflow for matching reference images in
references/reference-image-match.mdinvolves processing untrusted external data (user-provided images). However, the skill implements a structured 'Reference Analysis' phase that requires the agent to extract specific visual properties (silhouette, proportions, lighting, color palette), which effectively mitigates the risk of indirect prompt injection by focusing the agent on data extraction rather than instruction following.
Audit Metadata