Skip to content
Get packs

Robotics Idea Discovery

Guides an agent through a sim-first robotics research pipeline: survey, idea generation, novelty check, review.

Source Download skill
Repository
wanshuiyin/Auto-claude-code-research-in-sleep
Stars
18,482
Licence
MIT
Files
1 file

What it does

This skill makes the agent run a multi-step workflow that turns a broad robotics topic (such as manipulation or drones) into a short list of testable research ideas. It frames the problem around robot details like embodiment, benchmarks and sim-to-real constraints, then chains literature search, idea generation, novelty checking and critical review. The workflow favours simulation and offline data over real hardware, and it is written so the agent proceeds without waiting for answers at checkpoints unless the user stops it.

Great for

  • Robotics or embodied AI researchers looking for benchmark-grounded paper ideas
  • Students planning a thesis or project in robot learning who want a structured starting point
  • Teams wanting a simulation-first pilot plan before committing hardware time

Skip it if

  • Users without access to the sub-skills it calls (/research-lit, /idea-creator, /novelty-check, /research-review), since the pipeline depends on them and this file alone does not define them
  • Anyone expecting a verified novelty judgement: the novelty check relies on search and model review, which can miss prior work
  • Real-robot experiments run from this skill's output without human review, because it is designed to avoid hardware execution unless explicitly approved

Try asking

  • “robotics idea discovery for bimanual manipulation, sim only”
  • “机器人方向探索: drone navigation ideas targeting CoRL and RSS”
  • “Find sim2real research ideas for quadruped locomotion using public benchmarks”
  • “Survey embodied AI benchmarks like ManiSkill or LIBERO and suggest gaps”
Step-by-step guide: how to use Robotics Idea DiscoveryInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    idea-discovery-robot.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip idea-discovery-robot.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip idea-discovery-robot.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • State the embodiment, task and whether you have a simulator in the request, e.g. "drone navigation in Isaac Lab, no real robot", so Phase 0 needs fewer assumptions.
  • Override defaults inline, such as limiting pilot ideas to 2 or restricting targets to CoRL/RSS.
  • Review the checkpoint survey before letting it continue, since auto-proceed will choose options without you.

What's inside

  • SKILL.md

Redistributed unmodified from wanshuiyin/Auto-claude-code-research-in-sleep under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.