blender-video-recreate

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the npx skills add command to install domain-specific extensions (e.g., blender-scene-assembly, blender-character-rigging) from the full-aigc-skills organization. These downloads are part of the author's ecosystem for extending the agent's Blender-related capabilities.
  • [COMMAND_EXECUTION]: The skill executes shell commands via npx to manage dependencies and install the necessary Blender skills for the reconstruction workflow.
  • [INDIRECT_PROMPT_INJECTION]: The skill analyzes user-supplied reference videos, which serves as a potential ingestion point for untrusted data that could influence the generated prompts.md used for downstream video generation.
  • Ingestion points: Reference video files provided by the user in Step 1 of the workflow.
  • Boundary markers: The skill includes checks to verify if media can be decoded before interpretation begins and emphasizes recording per-shot facts independently.
  • Capability inventory: The skill performs local Blender scene mutations, rendering, and can optionally submit tasks to remote video generation providers like Jimeng and Seedance.
  • Sanitization: While no specific text-filtering is mentioned for video metadata, the skill mandates technical probing of file properties and requires explicit human-in-the-loop authorization for any network-based task submission or costs.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 02:22 PM