Skip to content
Get packs

Experiment Plan Builder

Turns a research idea into a claim-driven experiment roadmap with ablations, run order and compute budget.

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

What it does

Once installed, the agent takes a research proposal or method idea and produces a structured experiment plan. It defines the claims a paper must defend, the minimum evidence for each, the key experiment blocks (main results, ablations, simplicity and necessity checks), a staged run order with decision gates, and a compute and risk budget. The output is written to a markdown file in a refine-logs/ folder, so a researcher can follow it step by step.

Great for

  • Researchers or PhD students preparing paper experiments after their method is settled
  • ML engineers who need a realistic run order and GPU-hour estimate before starting training
  • Teams using the companion research-refine skill who want the next step in the chain

Skip it if

  • Early brainstorming when the method is not yet stable, since the plan assumes a fixed proposal and may produce false precision
  • Non-research projects, because the structure is built around paper claims and reviewer concerns
  • Anyone expecting the agent to actually run experiments: the skill only plans them, and its cost estimates are guesses the user must verify

Try asking

  • “Turn my refined proposal in refine-logs/FINAL_PROPOSAL.md into a detailed experiment plan with ablations”
  • “Create a claim-driven evaluation protocol and run order for my diffusion-based method, with a compute budget”
  • “Design an ablation matrix that tests whether my gain comes from extra parameters rather than the new mechanism”
Step-by-step guide: how to use Experiment Plan BuilderInstall, 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.

    experiment-plan.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip experiment-plan.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip experiment-plan.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

  • Run the research-refine step first so the problem anchor and dominant contribution are clearly defined
  • Ask for a shorter plan if your budget is small, since the skill caps primary claims at two and core blocks at five
  • Check the generated GPU-hour and timeline estimates against your own hardware before committing

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.