google-deepmind/robotics-antigravity-plugin

iteration

Staged execution and video-based debugging workflow for physical or MuJoCo simulated robots: splitting a robot manipulation task into stages, testing each stage with one script, reviewing rollout videos, and verifying physical completion after release and s…

ソースを見る

使い方

  1. インストールコマンドをコピーし、自分の環境へ導入する前にソースを確認します。
  2. 必要なツール、認証情報、対応環境を原文で確認します。
  3. 入力と成果物を指定し、出力を資料と照合します。
npx skills add https://github.com/google-deepmind/robotics-antigravity-plugin --skill iteration

開始前の確認

  • 導入にはモデル契約、API枠、ファイルへのアクセス許可は含まれません。
  • スター数や導入数は活動の指標であり、独立した品質・安全性試験ではありません。
リポジトリの原文

見出し、例、コード、表、リンク、参照画像を含む原文を表示しています。

Iterating on robot code

Work in stages

  • Split a long task into stages, one per sub-goal. For example, putting five

objects in a box is five stages. An action that already works reliably, such as a closed-loop pick-and-place, counts as one stage.

  • Write each stage as one script that observes, plans, acts, measures the

result, and prints it. Run it with a timeout, for example timeout 300 python3 stage_1.py.

  • After each stage, use the measured result to decide whether to move on. If

the stage failed, look at the trajectory to find where it went wrong and fix the code. Then measure the scene again and go back to the first stage whose goal no longer holds.

  • Do not put multi-line code in python3 -c. Write it to a file instead.

Get a second opinion when stuck

Use this when you can record video and start subagents.

  • When a stage fails repeatedly, give a separate subagent the video of the

last try (a video, not still images) and ask it to describe what happened and what went wrong. Send only the video, not your code or what you expected, so that its description is not biased.

  • If the stage still fails after that, give a separate subagent the videos,

your code, and the measured results, and ask what to change.

Done means measured

  • Decide success from measurements taken after the robot has let go and the

scene has settled, not from commanded poses or a printed "SUCCESS". Check every goal condition before you finish.

同じリポジトリから

関連する Skills

すべての Skills
google-deepmind
コミュニティ

blender-creation

Creates physically calibrated, articulated, PBR-textured 3D MuJoCo assets (MJCF model.xml, .mjz, mjSpec) from scratch using headless Blender (bpy, bmesh, blendermeshutils, assetbuildutils), convex collision synthesis, 10-scenario auditing, and Studio preview. Use only when asked to create, model, or build a 3D MuJoCo object from a URL, image, or text description, or when blender-creation is named. Don't use for non-MuJoCo Blender/Cycles/CAD modeling.

1時間の導入数
2
GitHub Stars
80
更新日
11月17日
google-deepmind
コミュニティ

fr3-duo-control

Franka Duo robot only (Franka Research 3 / FR3 Duo dual-arm cell with Robotiq 2F-85 grippers, ZED Mini and RealSense D405 cameras). Brings up, homes, and controls it via pylibfranka on hardware or in MuJoCo sim. Use only when the user names Franka Duo, FR3 Duo, or fr3duo.

1時間の導入数
2
GitHub Stars
80
更新日
11月17日
google-deepmind
コミュニティ

i2rt-yam-bringup

Checks and installs the I2RT YAM SDK (i2rt), configures Linux socketcan or macOS gsusb CAN bringup for I2RT YAM robot arms, and provides a wave smoke test (wave.py). Use only before bringing up or programming a physical or simulated I2RT YAM robot arm, or when asked to run a wave gesture on an I2RT YAM robot. Don't use for non-YAM CAN bus debugging (e.g. automotive OBD-II or industrial PLC CAN networks) or non-robotics tasks.

1時間の導入数
2
GitHub Stars
80
更新日
11月17日
google-deepmind
コミュニティ

i2rt-yam-ik

Solves 6-DOF inverse kinematics for the I2RT YAM and YAM Pro robot arm using Mink + DAQP on combined YAM + gripper MJCF. Use only when computing joint targets or local Cartesian moves on an I2RT YAM or YAM Pro robot arm. Don't use for non-YAM inverse kinematics, general quadratic programming (QP) solvers, or non-robotics tasks.

1時間の導入数
2
GitHub Stars
80
更新日
11月17日