add-task-env

Scaffold EmbodiChain task environments with Gym IDs and test stubs.

206|20|Updated Oct 24, 2025
One-click install
npx skills add https://github.com/DexForce/EmbodiChain --skill add-task-env
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: add-task-env
Source: https://github.com/DexForce/EmbodiChain/tree/main/skills/add-task-env
Command: npx skills add https://github.com/DexForce/EmbodiChain --skill add-task-env

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you quickly add a new task environment to EmbodiChain without manually figuring out the required file layout, registration hooks, and testing scaffolds.

Core Features & Use Cases

  • Task categorization: Choose the right category mapping to EmbodiChain’s task folder structure (tableware, rl, special).
  • Environment scaffolding: Generate a task Python module with the correct class shape, @register_env decorator usage, and extension points for expert demos (create_demo_action_list) or RL tasks (get_reward, compute_task_state, check_truncated).
  • System integration: Update tasks/__init__.py exports so the new environment is discoverable by the framework.
  • Quality gate: Add a test stub and run formatting with black to match repository conventions.

Quick Start

Use the add-task-env skill to scaffold a new task environment named pick_place for the category rl with Gym ID PickPlace-v1, then add it to tasks/init.py and create the corresponding tests stub.

Frequently Asked Questions about add-task-env

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I scaffold a new gym environment for reinforcement learning in EmbodiChain?

Scaffolding a new gym environment requires selecting a task category and snake_case name, then generating a Python module with reward and task state extension points using the EmbodiedEnv conventions. The environment is registered with a unique Gym ID and exported in tasks/__init__.py.

What is the correct way to register a unique Gym ID for an embodied AI task?

Registering a unique Gym ID involves applying the @register_env decorator within the new task Python module. The module must follow the EmbodiedEnvCfg conventions and be exported through tasks/__init__.py to become discoverable by the EmbodiChain framework.

Can I create expert demonstration tasks that define demo action sequences in EmbodiChain?

Yes, creating expert demonstration tasks is supported by defining the create_demo_action_list extension point within the scaffolded Python module. This follows the EmbodiedEnv conventions alongside RL task methods like get_reward and compute_task_state.

What task categories are available when adding a new reinforcement learning environment?

Available task categories map directly to EmbodiChain's folder structure: tableware, rl, and special. Selecting the correct category determines where the Python module is created within the tasks directory hierarchy.

Do I need to add test stubs when scaffolding new embodied AI task environments?

Yes, adding a test stub is required to pass the quality gate and match repository conventions. The scaffolding process also runs formatting with black to ensure the new environment module adheres to the project's code style standards.

Why are my newly created EmbodiChain task environments not discoverable by the framework?

Environments become discoverable only when properly exported through tasks/__init__.py. If the @register_env decorator is missing or the unique Gym ID is not registered within the Python module, the EmbodiChain framework cannot locate the new task.