robotwin-instruction-to-env

What problem does it solve?

This Skill turns a task name and task description into a RoboTwin environment, then renders and checks the first frame so the result is directly usable. It inspects the local repo, reuses nearby env and asset patterns, writes conservative scene setup code, and iterates until the frame is reasonable. It now also uses local asset pose priors so default orientations are grounded instead of guessed.

Core Features & Use Cases

• Normalize the user request into concrete scene requirements, including task name, task description, main objects, supports, containers, layout relations, embodiment constraints, camera/background constraints, and safe defaults for anything unspecified.
• Discover the target RoboTwin-like repo and validate compatibility, reuse nearby env patterns, load pose priors, generate envs/<task>.py, render the first frame, and review against a first-frame checklist.
• Render and validate a deterministic first frame using the repo's render path, with guidance from local pose manuals and repository conventions to ensure a stable initialization for downstream testing or demonstrations.

Quick Start

Provide a task name and description; the skill will build a RoboTwin env and render the first frame.