[Feature]: Skill category: Embodiment/Physical world/World modelling for robotics

[clayrosenthal]

Problem Statement

I didn't see any top-level category or sub-skills for physical tasks. There aren't many agents currently that can do physical work, but could make sense to plan for the future and assign a category for it. Not sure if it should be it's own category, but something like Genie 3 from Google's world modelling is also a powerful skill I could see agent's leveraging soon.

Proposed Solution

Make another section called Embodiment or Physical Intelligence or World Modelling and add skills for like:

• Bipedal motion
• 3D navigation
• Quadruped motion
• Wheeled motion
• One arm manipulation
  • pinch end effector
  • hand end effector
  • suction cup end effector
• Two arm manipulation
  • ... same end effectors as one arm
• World Modeling
• etc...

Alternatives Considered

Add these skills to an existing category, maybe Multi-Modal?

Additional Context

Agentic robots are certainly possible sooner or later, would make sense to add those as options.

Checklist

• I have read the contributing guidelines.
• I have verified this does not duplicate an existing feature request.

[muscariello]

Hi @clayrosenthal

I would add new categories like you suggest. I was working on expanding the taxonomy already for other applications
and I could use that to show how to expand the taxonomy.