The Agent Skills Directory

[SAFE]: The skill uses the standard Blender Python API (bpy) to perform rendering tasks. All operations, such as modifying scene properties and creating lights, are within the expected functional scope of a Blender rendering utility.
[INDIRECT_PROMPT_INJECTION]: The skill's architecture for processing user rendering requests was evaluated for potential vulnerabilities.
Ingestion points: Natural language triggers such as "render", "lighting", and "EEVEE" in SKILL.md.
Boundary markers: The skill provides specific Python scripts and workflows, acting as a structured guide rather than raw user input interpolation.
Capability inventory: Blender API access is limited to scene modification and property management; file path settings use project-local relative paths.
Sanitization: Input is restricted to valid Blender property values through hardcoded script templates.
[EXTERNAL_DOWNLOADS]: The skill previously included networking components (curl, requests) which have been explicitly removed in version 2.0.0 to reduce the attack surface. Assets such as HDRI textures are configured to load from local relative paths (SKILL.md, ADVANCED_RENDERING.md).

blender-rendering