The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied 3D scene data and generates JSON reports for agent consumption.
Ingestion points: The scripts/audit_material_uv.py script reads metadata such as object names, material names, and file paths directly from the active Blender scene.
Boundary markers: The script output is raw JSON and lacks boundary markers to instruct the agent to ignore potentially malicious instructions embedded in scene data.
Capability inventory: The skill utilizes blender-python-execution for script execution, providing access to the local filesystem and Blender API.
Sanitization: No sanitization is performed on strings derived from the Blender scene before they are included in the JSON audit report.
[DYNAMIC_EXECUTION]: The workflow relies on generating and executing Python code at runtime. The documentation in SKILL.md describes a pattern for dynamically assembling the audit script by prepending configuration variables before execution. Additionally, references/bpy-recipes.md provides code templates intended for dynamic instantiation and execution via the Blender Python environment.
[EXTERNAL_DOWNLOADS]: The instructions recommend using an external bootstrap skill to download and install additional skills such as blender-uv-texturing, which introduces a dependency on external repositories for extended functionality.

texture-uv-material-workflow