Skip to content
Get packs

ResearchClaw Autonomous Research Pipeline

Runs a 23-stage pipeline that turns a research topic into a draft paper with experiments and charts

Source Download skill
Repository
aiming-lab/AutoResearchClaw
Stars
14,802
Licence
MIT
Files
1 file

What it does

Once installed, the agent can take a research topic and drive ResearchClaw's multi-stage pipeline end to end: literature review, hypothesis generation, experiment design, code generation and execution, result analysis, paper writing, peer review and export. It sets up a config file with your LLM API key, runs the pipeline via a command-line tool or Python code, and reports where outputs such as a paper draft, results tables and charts were saved. The pipeline itself is an external project; this skill tells the agent how to operate it.

Great for

  • Researchers or students who want a first-draft paper and experiment scaffold on a topic quickly
  • Developers prototyping automated research workflows who already have ResearchClaw installed
  • Data scientists who want simulated or sandboxed experiment runs with generated metrics and charts

Skip it if

  • Producing publication-ready research: the skill itself describes LLM-generated results and papers, which need expert verification before any use (risk of fabricated or unverified findings)
  • Running on a machine with sensitive data or credentials without review, because the pipeline can execute generated code locally (sandbox mode) or on a remote server (ssh_remote mode)
  • Users without an LLM API key, since the pipeline depends on one and will fail to connect otherwise

Try asking

  • “Research the effect of spaced repetition on language learning and draft a paper”
  • “Write a paper about sparse attention methods using the ResearchClaw pipeline”
  • “Run ResearchClaw on my config.yaml with the topic 'graph neural networks for fraud detection' and auto-approve the gates”
  • “Resume the ResearchClaw run from the PAPER_OUTLINE stage in my existing artifacts folder”
Step-by-step guide: how to use ResearchClaw Autonomous Research 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.

    researchclaw.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

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

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip researchclaw.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

  • Start with experiment.mode set to simulated to see the full workflow without executing any code, then move to sandbox after reviewing the generated experiment.py
  • Review the config file and generated experiment code before running, and keep API keys in environment variables rather than in config.yaml
  • Treat the output paper draft as a starting point and verify every cited result and claim yourself

What's inside

  • SKILL.md

Redistributed unmodified from aiming-lab/AutoResearchClaw 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.