Skip to content
Get packs

Research Review via Claude

Get multi-round adversarial peer review of research from Claude through a local MCP bridge

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

What it does

Once installed, a Codex agent gathers your research notes, paper drafts and experiment history, then sends them to Claude through a local reviewer bridge (the claude-review MCP server). Claude is asked to act like a tough NeurIPS/ICML reviewer, listing logical gaps, missing experiments and weak claims. The agent then runs follow-up rounds with Claude, and saves the whole exchange and conclusions to a review document in your project.

Great for

  • Researchers who want a critical second opinion on a paper draft or result before submitting
  • Codex users who want Claude, a different model family, to review work done by a Codex agent
  • Students preparing for peer review who want a mock reviewer with scores

Skip it if

  • Final security or compliance audits: the skill tells the reviewer to assume a cooperating operator and not to look for malicious users, so it is not a security review tool
  • Confidential or unpublished work you are not allowed to send to a third-party model, because your project files and prompts are sent to Claude
  • Users who do not run Codex or Claude Code, since it depends on a Codex-specific setup and an MCP server

Try asking

  • “Review my research draft and tell me where the logic is weakest”
  • “Get external review of my experimental results and suggest the minimum extra experiment to satisfy the reviewers”
  • “Write a mock NeurIPS review with scores for my paper outline”
  • “Build a results-to-claims matrix for possible outcomes of my next experiment”
Step-by-step guide: how to use Research Review via ClaudeInstall, 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.

    research-review.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

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

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip research-review.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

  • Give the agent clear context first (README, STORY.md, paper draft, notes) so the reviewer has the full picture
  • Ask targeted follow-ups on the most actionable criticisms rather than accepting every point
  • Check the saved review document yourself, since the skill records the reviewer's conclusions as accepted

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.