Run Agentic Environment Generation#

Complete the shared Prerequisites before running this workflow.

The Arena environment generation agent infers an ArenaEnvGraphSpec YAML from a user prompt. The agent runs in two modes:

  • GUI runner — a browser live editor. Generate from a prompt, then edit, visualize, and simulation-preview the spec in the same session.

  • CLI runner — a one-shot, non-interactive pipeline. It writes the YAML and editing can be done manually in a text editor. Use it for scripted or batch generation.

Name the background prims the task depends on — the fridge and the floor — so the agent emits them as object_references instead of spawning new assets, and state the openness the door has to reach to complete the task. That is, the task is to open the fridge door to the 0.2 openness threshold in the OpenDoorTask.

Start the live editor and open http://localhost:8501 in a browser:

python isaaclab_arena_examples/agentic_environment_generation/gui_runner.py

In the Generate from prompt panel, enter the prompt and click Generate spec:

There is a floor and a fridge in the lightwheel_kitchen_one_wall_coastal kitchen.
DROID is on the floor, next to the fridge with 0.1 meter distance and facing
it. DROID opens the fridge door to the 0.2 openness threshold.
GUI runner view of the environment graph spec.

The prompt generates the floor and fridge references, DROID placement relations, and the fridge-opening task.#

Run the runner in resolve mode:

python isaaclab_arena_examples/agentic_environment_generation/cli_runner.py \
   --mode resolve \
   --prompt "There is a floor and a fridge in the lightwheel_kitchen_one_wall_coastal kitchen. DROID is on the floor, next to the fridge with 0.1 meter distance and facing it. DROID opens the fridge door to the 0.2 openness threshold."

The runner prints the resolved graph and writes <env_name>.yaml under isaaclab_arena_environments/agent_generated/.