blender-syntax-materials

Warn

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill's metadata contains deceptive authorship information. The SKILL.md frontmatter identifies the author as 'OpenAEC-Foundation', which does not match the known author 'Impertio-Studio' or its established naming patterns. This misleading metadata can lead to incorrect trust assessments by users or agents.- [PROMPT_INJECTION]: The skill provides patterns that process untrusted external data, creating a surface for indirect prompt injection.
  • Ingestion points: The skill accepts external file paths through variables like 'texture_dir' in Example 2 (references/examples.md) and reads attributes from active Blender objects.
  • Boundary markers: No delimiters or explicit instructions are provided to the agent to treat processed data as untrusted or to ignore embedded commands.
  • Capability inventory: The skill utilizes the 'bpy' API and 'os' module to perform file system operations (bpy.data.images.load) and manipulate the Blender scene and object data.
  • Sanitization: The provided code patterns do not include logic for sanitizing, validating, or escaping input paths or mesh data content before use.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 18, 2026, 02:17 PM
Security Audit — agent-trust-hub — blender-syntax-materials