Skip to main content
skill CX Codex

robodojo-return-to-start — Claude Code skill for arm reset

The robodojo-return-to-start Claude Code skill applies RoboDojo’s finishing posture after the general return procedure. It restores used empty arms to their initial poses and checks their clearance from the result.

MIT ★ 488 updated 2026-11-15

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add Open-RUA/uni-robot-agent --skill robodojo-return-to-start -a claude-code
Open on GitHub

What robodojo-return-to-start does

The skill sets the RoboDojo arm finish after the general return procedure. It reads each arm’s starting control_pose from step=0. When the task requires no continued contact, each used empty arm returns to that pose under the task’s and system prompt’s finishing requirements. The skill checks position and orientation, then confirms that neither arm disturbs the result.

The general procedure uses general/gripper/return-to-start/SKILL.md. RoboDojo’s native return threshold is measured at link6, not at the grasp site; an orientation error can shift link6 even when the grasp-site position matches. The skill provides no universal numeric tolerance. It follows the threshold declared by the current task and does not convert it into a grasp-site position tolerance.

What it helps with

  • return an empty arm to its starting pose after a task
  • check both arms’ position and orientation
  • keep the arms clear of the completed result
  • apply a task’s native link6 return threshold

Example prompts for the agent

  • After the task, return both empty arms to their starting poses and check that they leave the result untouched.
  • Apply the RoboDojo return procedure and compare link6 with the task threshold.
  • Read the initial control_pose values from step=0 and finish under the system requirements.

How to install robodojo-return-to-start

1Skills CLI (Vercel)

npx skills add Open-RUA/uni-robot-agent --skill robodojo-return-to-start -a claude-code

2Manually: put the files in place

~/.claude/skills/return-to-start/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/return-to-start/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about robodojo-return-to-start

How do I install robodojo-return-to-start in Claude Code?

Run npx skills add Open-RUA/uni-robot-agent --skill robodojo-return-to-start -a claude-code. The skill file is at ~/.claude/skills/return-to-start/SKILL.md.

What does robodojo-return-to-start do?

It restores used empty RoboDojo arms to their initial poses after the general return procedure, then checks their orientation, position, and clearance from the result.

Which return threshold does robodojo-return-to-start use?

The native threshold is measured at link6, rather than at the grasp site. The skill uses the current task’s declared threshold and sets no universal numeric tolerance.

Source text

SKILL.md · MIT
---
name: robodojo-return-to-start
description: Apply RoboDojo finishing-posture and native link6 return semantics after the general return procedure.
---

# Apply the RoboDojo return reference

Use `general/gripper/return-to-start/SKILL.md` for release, clearance and pose
verification. Read each arm's initial grasp-site control_pose from step=0.
When no continued contact is required, return each used empty arm to that pose,
following the task and system prompt's finishing requirements. Verify position
and orientation, and check both arms are clear without disturbing the result.

Official native return-position thresholds are
measured at link6, not at the grasp site; orientation error can displace link6
even when the site position matches. Restore the full initial pose accurately.
Do not reinterpret a native link6 position threshold as a grasp-site position
tolerance or add stricter task requirements. Follow the current task's declared
thresholds; this skill does not supply a universal numeric return tolerance.

© Open-RUA · MIT · Open-RUA/uni-robot-agent

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set41

The whole set →
hybrid_decision The hybrid_decision prompt for Claude Code reviews an agent’s proposed action against observations and execution history. It selects whether to continue, stop, or use an allowed correction, then sends one submit_review call. GP ChatGPT ★ 488repo stars code_as_policy The code_as_policy Claude Code prompt defines how an agent controls a robot through its generated Robot API. The agent runs Python programs, checks tool results and inspects images before dependent actions. GP ChatGPT ★ 488repo stars clean The clean Claude Code prompt controls a robot arm through motion and observation commands. It guides instruction execution and checks results between actions. GP ChatGPT ★ 488repo stars system The system Claude Code prompt guides control of a dual-arm ARX X5 robot in a live RoboDojo episode. The agent consults skills, observes the scene, and checks action results before proceeding. GP ChatGPT ★ 488repo stars deformable-material The deformable-material Claude Code skill moves selected regions of flexible material and checks their shape after release. It applies when deformation makes rigid-object poses or earlier contact coordinates unreliable. CX Codex +1 ★ 488repo stars contents-transfer contents-transfer — a skill for Claude Code that transfers loose contents from a held vessel into a receiver. It tracks outlet alignment, flow, and what remains. CX Codex +1 ★ 488repo stars moving-target-control The moving-target-control Claude Code skill plans contact with moving physical objects and checks whether a grasp holds. It activates when a target or receiver may move during approach or transfer. CX Codex +1 ★ 488repo stars return-to-start return-to-start — a skill for Claude Code and Codex guides robotic arms back to their measured starting poses. Verifies position and orientation tolerances before ending the movement. CX Codex +1 ★ 488repo stars

Project sponsors

Companies that keep this analytics open