Preview
Sim checkpoint replay (play.sh). Not the ONNX runtime path, not a real robot, not an official Pollen policy.
polite-bow
Bows from a two-foot stand and comes back up. Community example from the microduck-plugin loop. Not an official Pollen policy. Weights live here; they are not bundled in the plugin git.
A episodic policy for the microduck (61-D observation, 14 actions, 50 Hz). Runs 4.0 s and returns itself to a standing pose.
Do not start --repo this graph as the standing body. Start a walk/stand ONNX, pull --as polite_bow, then do polite_bow.
Agent plugin / skill
git clone https://github.com/acnlabs/microduck-plugin
# Point an Agent Plugins client at that folder (the one with plugin.json).
Skill-only clients:
npx skills add acnlabs/microduck-plugin@microduck-skill -g -y
Then the agent runs localhost control:
control.sh start --repo neil-jo/microduck-walk --detach
control.sh pull neil-jo/microduck-polite-bow --as polite_bow
control.sh do polite_bow
ClawHub listing for microduck-skill is still pending review.
If you own a Microduck
sudo robotctl policy add polite-bow neil-jo/microduck-polite-bow
robotctl robot do polite-bow
The observation normalizer is baked into policy.onnx; feed raw observations.
manifest.json follows schema 2 of the microduck policy manifest (docs/policy-manifest.md in the daemon repo).
Training
- repo:
pollen-robotics/microduck_rl - branch:
develop - commit:
29e887ecf - exported from a checkout with a local polite-bow env (
Mjlab-PoliteBow-Flat-MicroDuck), not yet upstream