blender-asset-discovery-planner

Pass

Audited by Gen Agent Trust Hub on Oct 24, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external project briefs and asset descriptions which could theoretically contain malicious instructions. This risk is effectively mitigated by the skill's lack of hazardous capabilities, as its output is restricted to plain markdown text.
  • Ingestion points: External 3D project briefs and user-supplied asset library lists as defined in SKILL.md.
  • Boundary markers: Not explicitly implemented to isolate untrusted data.
  • Capability inventory: Restricted to markdown text output; subprocess calls, file writes, and network operations are explicitly prohibited by the skill's operational laws.
  • Sanitization: No specific input sanitization identified.
  • [EXTERNAL_DOWNLOADS]: The skill mentions various external asset sources and community tools (e.g., Poly Haven, Sketchfab, mcp-for-blender). These references are for documentation and planning purposes only. The skill instructions and validation checklists explicitly forbid the downloading, fetching, or installation of any external assets or code during operation.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 24, 2026, 07:02 PM