The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill generates Python code in strings and executes them using the execute_blender_code tool. This is the core mechanism for automation but represents a dynamic code execution surface.- [EXTERNAL_DOWNLOADS]: Reference files reference/ADVANCED_SCULPTING.md and reference/TERRAIN_GENERATION.md include import requests, which provides the capability for network operations, although no specific malicious URLs are targeted.- [INDIRECT_PROMPT_INJECTION]: The skill ingests heightmap images without sanitization or boundary markers, using them to drive vertex displacement.
Ingestion point: bpy.data.images.load in reference/TERRAIN_GENERATION.md.
Boundary markers: Absent; no instructions are provided to the agent to treat external asset data as untrusted.
Capability inventory: The skill has access to the execute_blender_code tool for Python execution and performs file system writes via image.save().
Sanitization: Absent; pixel data from external files is directly used to modify mesh geometry coordinates.

blender-sculpting