Generate Python robot manipulation task definitions for the RoboLab benchmark framework.
npx skills add https://github.com/NVlabs/RoboLab --skill robolab-taskgen
Please help me install this Agent Skill. Skill: robolab-taskgen Source: https://github.com/NVlabs/RoboLab/tree/main/skills/robolab-taskgen Command: npx skills add https://github.com/NVlabs/RoboLab --skill robolab-taskgen
This skill removes the complexity of manually authoring robot manipulation tasks by automating the creation of Python task definitions, success conditions, and instruction variants from natural language.
Use the robolab-taskgen skill to create a new task file for picking up a red block and placing it on a blue block in the workspace.
High-intent search queries and answers about installing and using this skill.
To generate robot manipulation tasks for Isaac Lab, provide a natural language description of your scene and goals. The skill automatically outputs structured Python task classes, subtask decomposition, and mapped termination criteria like object_in_container.
Creating a manipulation benchmark task requires valid USD scene assets and access to the robolab.core.task module. These inputs ensure environment compatibility when the skill generates the Python task definitions and success conditions.
You can map natural language goals to predefined termination functions in robot simulations automatically. The skill interprets your text input and links it to specific termination criteria like stacked or object_in_container within the generated Python code.
Robolab-taskgen supports subtask decomposition for complex pick-and-place benchmark tasks. It breaks down the natural language input into structured subtasks and formats all required termination logic within the generated Python task class.
Manually authoring robot manipulation task definitions is complex due to formatting termination criteria and subtask logic. You can automate this by using natural language input to generate structured Python task classes compatible with the RoboLab benchmark framework.
NVIDIA-OmniversePrint USD scene hierarchy and prim metadata from readable input files.
hkqaiGenerate Isaac Lab simulation scripts from YAML scene descriptions.
Kling0012Create and train AI learning plugins with AgentDB reinforcement algorithms.
sourcegraphAudit benchmark suites against the Agent Benchmark Criteria framework.
