- Home
- Skills
- Agents & Orchestration
- Humanoid Whole-Body Control & Teleop
Humanoid Whole-Body Control & Teleop
Humanoid robot engineering for whole-body control, balance, motion retargeting, and safety-critical hardware deployment.
$49
Works with the AI Agents you already use
Humanoid Whole-Body Control & Teleop
Example session with this skill installed
Set up a whole-body control task for a Unitree G1 to pick a 2kg box from a 75cm table and place it on a 1.2m shelf. Include safety constraints and the retargeting logic for VR teleoperation.
- Read your context and instructions
- Compiled the humanoid whole-body control
I have structured the whole-body control task for the Unitree G1, including the QP task hierarchy, safety bounds, and the motion retargeting configuration for your VR data collection.
Humanoid Manipulation Task: Box Transfer
| Component | Specification | Constraints |
|---|---|---|
| Primary Task | End-effector SE(3) tracking | Max velocity: 0.5 m/s |
| Postural Task | Joint space regularization | Weighted toward singularity avoidance |
| Balance Task | Centroidal Momentum / CoM | Keep CoM within 80% of support polygon |
| Safety Limits | Torque caps & Workspace | Z-limit: 0.2m to 1.8m; Torque: 40Nm (arms) |
| Retargeting | Human-to-G1 Mapping | 50ms latency target; Exponential smoothing |
# Unitree G1 Whole-Body QP Task Configuration
task_config = {
"robot": "unitree_g1",
"tasks": {
"contact": {"type": "fixed", "link": ["left_foot", "right_foot"]},
"balance": {"type": "com_tracking", "weight": 1000},
"manipulation": {
"type": "se3_tracking",
"link": "right_hand",
"goal": [0.6, 0.0, 0.75], # Table height
"weight": 500
}
},
"safety": {
"collision_avoidance": True,
"self_collision_buffer": 0.02, # meters
"fall_detection": {
"pitch_threshold": 0.3, # rad
"action": "active_damping"
}
}
}
Next steps
- Run the task in simulation (MuJoCo/Isaac) to verify the QP solver convergence at 500Hz.
- Calibrate the VR hand-tracking offset relative to the G1 shoulder base frame.
- Perform the first 10 test episodes on the safety gantry before attempting free-standing placement.
Connects securely to your tools. The creator never sees your data.
About this skill
Balance theory, WBC QP, motion retargeting, VLA data collection.
Engineer humanoid robot behaviours (Unitree G1/H1, Fourier, Booster, open-source humanoids) - whole-body control (QP/TSID, centroidal MPC), balance and walking, motion retargeting from human motion capture, VR teleoperation for data collection, imitation learning / VLA policies for manipulation, and staged hardware testing with a safety gantry. Use when the user programs a humanoid to walk, balance, imitate human motions, manipulate objects with both arms, or collects demonstrations for robot learning.
Includes an EXPERT.md reference (exact parameters, thresholds, failure diagnosis).
Acceptable use: civil and commercial applications only. Any military, paramilitary or weapon-related use is prohibited, as stated in the skill itself.
Requires external simulation and physical safety hardware. Does not include low-level proprietary motor firmware. Scope is limited to civil and commercial applications only.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 28 days ago