The Agent Skills Directory

[PERSISTENCE]: The INSTALLATION.md file provides instructions for users to modify their local ~/.nuke/menu.py configuration to integrate the tool into the Nuke UI. This is the standard method for installing persistent tools and menu items within the Nuke ecosystem.
[COMMAND_EXECUTION]: The skill utilizes Nuke's internal Python interpreter to execute local script files via execfile and programmatically constructs complex node hierarchies using the nuke.nodes module to automate tiling and merging workflows.
[DYNAMIC_EXECUTION]: The auto_tile_processor.py script dynamically generates Nuke Expression nodes containing mathematical formulas (smoothstep) to handle gradient blending between tiles at render time.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The tool processes image data from a selected Nuke input_node (e.g., a Read node), which serves as the primary data ingestion point.
Boundary markers: As the tool operates on image data and node structures rather than natural language prompts, no explicit boundary markers or delimiters are employed.
Capability inventory: The skill has the capability to create, delete, and modify Nuke nodes and execute Python code within the host application environment.
Sanitization: The implementation includes input validation for tile_size and overlap parameters, checking for valid integer types and logical range limits.

nuke-tiling-tool