DM Control finger spin feedback policy: robotics skill example

DM Control finger spin feedback policy. Bounded native-control simulation with privileged state; not a trained policy or hardware deployment.

Markdown | JSON example and evidence | Canonical contract | All examples

Task and environment

dm-control | 1.0.34 | finger | privileged_state_bounded_feedback | linux | 3.10

Example input

{
  "seed": 0
}

Retrieve without installation

curl -fsSL 'https://api.superfastrobotics.com/v1/skills/dmcontrol-finger-spin?hub_commit=ede75c00118e9808e5099da43c26f4619cc3bc37&record_version=1.0.0&include=experience'

Inspect source, license, environment and execution evidence before preparing. Preparation and execution are separate actions.

Success condition

All declared goal and finite-state checks pass; exit 0 with task_complete. Mean native task reward over the final 100 control steps >= 0.6 with finite state and nonzero control. Episode rewards are not task-agnostic accuracy.

Recorded evidence

import_verified: 0/3 frozen cases completed; retained failures are not promoted.

Source commit: 3831fa523c160fe653d83a829d0a0c59715f5725. Record version: 1.0.0. This page performs no execution.