Skip to content
Get packs

Idea Discovery Pipeline

Chains literature survey, idea brainstorming, novelty checks and reviews into one research workflow

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

What it does

This skill tells an AI agent to run a multi-step research workflow: survey the literature on a topic, brainstorm and rank candidate ideas, check whether they are novel, get critical feedback from a second reviewer model, and refine the best idea into a proposal and experiment plan. It writes the results into a report folder (idea-stage/IDEA_REPORT.md and refine-logs/ files). It is designed to run mostly automatically, chaining other sub-skills together.

Great for

  • Researchers who want a structured way to go from a broad research direction to a few testable ideas
  • Students or ML engineers starting a new project who need a literature-backed starting point
  • Teams that want an audit trail of how an idea was validated

Skip it if

  • Quick one-off brainstorming, since the full pipeline is long and makes many tool calls and external requests
  • Users without the referenced sub-skills (research-lit, idea-creator, novelty-check, research-review, research-refine-pipeline) installed, because the workflow depends on them
  • Anyone expecting verified truth: novelty and review steps are model judgements and can be wrong

Try asking

  • “Run the idea discovery pipeline on 'efficient fine-tuning for small language models'”
  • “Find research directions from zero on 'graph neural networks for drug discovery'”
  • “Use idea discovery with reference paper https://arxiv.org/abs/2406.04329 to improve an existing method”
  • “Run the complete idea exploration workflow in compact mode for a short-context model”
Step-by-step guide: how to use Idea Discovery PipelineInstall, 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.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

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

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip idea-discovery.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

  • Set AUTO_PROCEED deliberately: true runs checkpoints without pausing, false stops at each choice for your input
  • Provide a reference paper or base repo when you want to improve existing work rather than start from scratch
  • Review IDEA_REPORT.md yourself before spending GPU time on any pilot experiment

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.