The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation suggests an MCP server integration using the command npx -y @ahujasid/blender-mcp. This involves downloading and executing code from an unverified third-party developer on the NPM registry.
[REMOTE_CODE_EXECUTION]: The skill facilitates the execution of arbitrary Python code within the Blender process using flags such as --python and --python-expr, as well as a dedicated blender_execute_script tool. This creates a significant surface for executing dynamically generated code.
[COMMAND_EXECUTION]: The skill relies on executing the blender binary via shell commands for rendering, scene management, and batch processing.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its ingestion of external script data and 3D model files.
Ingestion points: Scripts processed by the blender_execute_script tool and model files loaded via blender_import_model.
Boundary markers: No delimiters or instructions to ignore embedded commands are present in the provided templates.
Capability inventory: The skill has access to the bpy API and system modules like os and glob for file operations.
Sanitization: There is no evidence of input validation or script sanitization before execution.

blender