SkillFoxx

Open ScholarPeer skill

Skills and commands for agents run a seven-step ScholarPeer review of a paper and write the result for a venue.

  • 52 ★ on the repository
  • Last commit
  • MIT
Skill

Extended access

Why this level

  • The `curl | bash` installer writes files and agent configuration into the launch directory.
  • The retrieval and onboarding steps query external literature sources and conference sites.
All reasons and checks

amirkiarafiei/open-scholar-peerAuthor:amirkiarafiei

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add skills/open-scholarpeer

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • The `curl | bash` installer writes files and agent configuration into the launch directory.
  • The retrieval and onboarding steps query external literature sources and conference sites.

This is third-party code. Review the repository files before installing.

What it does

Open ScholarPeer moves the seven-agent process from the ScholarPeer multi-agent review paper into a set of skills, slash commands and MCP tools installed into the project directory. The `/open-scholar-peer` command guides the steps: onboarding detects the venue and gathers review guidelines, followed by a paper summary, three-round literature retrieval, a historical narrative, an audit of missing baselines and datasets, criterion Q&A and the final review. Every artifact is saved as plain markdown in `.brain/raw/` and `.brain/review/`, so the reasoning can be audited. The installer creates `.brain/`, the agent configuration and a root `.mcp.json` or `AGENTS.md` file.

Who it is for. Researchers and reviewers who need a structured review draft for a paper with a literature check.

Good fit when

  • You need to prepare a review of a paper for a conference and onboarding fetches the venue rules.
  • You need to see which works the authors did not compare against: `/4-osp-baseline-scout` finds missing baselines and datasets.
  • The paper must be split into claims, method and evidence with `/1-osp-summary`.
  • A reviewer needs the chronological context of the field before reading: `/3-osp-historian`.
  • You want questions for each scoring criterion, with the count configurable (two by default).

Not a fit when

  • The review must be written entirely by a person with no model help: the tool generates a draft.
  • You only need a spelling or formatting check of the paper.
  • The agent has no access to literature search: the retrieval step uses live sources.

Example request

Review `paper.pdf`: detect the venue, summarize it, find missing baselines and prepare the final review under the venue rules.

Limitations

This is a community implementation of the ScholarPeer paper; the paper authors code is not used. The installer runs through `curl | bash` and writes files into the launch directory, so run it in the paper folder and read the script first. The literature step queries external sources (`arxiv`, `openalex`) and onboarding fetches conference guidelines. Review quality depends on the model and the result must be checked by you. The last commit was in November 2026.

How to disable. Delete the `.brain/` directory, the root `.mcp.json` or the `AGENTS.md` block and the skill and command folders the installer created in `.claude/`, `.cursor/` and other agent directories.

FAQ

How do I install Open ScholarPeer?

From the paper folder run `curl -sSL https://raw.githubusercontent.com/amirkiarafiei/open-scholar-peer/main/install.sh | bash`. Or clone the repository and run `bash install.sh --dir /path/to/paper`.

Which steps does the review pass?

Seven: onboarding, summary, literature, historian, baseline scout, Q&A and final review. Commands `/0-osp-onboarding` through `/6-osp-review`.

Where are the results kept?

In `.brain/raw/` and `.brain/review/` as markdown. Every artifact can be opened and checked.

Which agents are supported?

The installer configures Claude Code, Cursor and other agents; pick them with `--tool claude,cursor`. Literature sources are set with `--sources arxiv,openalex`.

What gets installed on my machine?

Only into the launch directory: `.brain/`, agent configuration with slash commands, skills and MCP, and a root `.mcp.json` or `AGENTS.md` file.

Discussion

No questions or reviews yet.

Skills

A skill set for academic work: paper review, pre-analysis plan check, grant proposal review, code audit and slide decks.

SkillExtended access560

A Claude Code skill for scientific literature reviews: an interview, Scopus and OpenAlex search, DOI checks and a final docx file

SkillExtended access371

A library of 280+ research skills: literature search, experiment design, analysis, reference checks and paper preparation

SkillExtended access124

A marketplace of five Claude Code plugins for researchers: literature search, research design, project memory and manuscript work

PluginBroad access655

A skill for master and PhD students that runs research through novelty and reality gates, from review to claim checks

SkillMinimal access71

The paper-spine skill takes a topic, materials and data through a scientific paper: literature, text, figures, review and Word, LaTeX and PDF files

SkillBroad access6.3K

README in short

The project assembles seven review steps into a set of skills, commands and MCP tools for agents. Progress and results are kept in markdown.

README badge

Original SKILL.md

---
name: open-scholar-peer
description: "Open ScholarPeer dispatcher — show review status and route to the next command"
---

# /open-scholar-peer — Stateless Dispatcher

This command is **always available**. It reads `.brain/session.json` and tells the user where they are in the review and which command to run next. It never executes review work itself.

## Activation

Invoke the `osp-orchestrator` skill.

## Steps

1. **Read `.brain/session.json`.** If it is missing, OSP is not initialised here. Two cases, and
   neither should send the user hunting for a script that is not in their project:
   - `.brain/` is absent as well — have them re-run the installer from this directory; it merges with
     their existing config rather than replacing it.
   - `.brain/` exists but the file does not — write it from the v2 schema yourself, then carry on.

2. **Print the status block** defined in `.agents/defaults/phase_block_template.md`, labelled
   `OPEN SCHOLARPEER`. The rail is the whole point here — it answers "where am I" with no words.
   Values:

       STEPS    onboard·summary·lit·history·baseline·qa·review
       VENUE    <name>   ·   <N> criteria
       PAPER    <path, or "not yet loaded">
       NOTE     literature stopped at round 2 of 3, by choice
       NEXT     /3-osp-historian   the narrative  (recommended)
                /5-osp-qa           skip ahead

   `NOTE` lists every phase whose `status` is `skipped`, plus any round or criterion left short;
   drop the label when there are none. `NEXT` recommends first, then offers.

3. **Route based on `resume_from`:**

   | `resume_from` | Tell user to run | Description |
   |---|---|---|
   | `onboarding` | `/0-osp-onboarding` | Set venue, locate paper, scaffold criteria |
   | `summary` | `/1-osp-summary` | Internal Compression — claims/method/evidence |
   | `literature` | `/2-osp-literature` | External retrieval, up to 3 rounds — the user chooses |
   | `historian` | `/3-osp-historian` | Build the chronological domain narrative |
   | `baseline_scout` | `/4-osp-baseline-scout` | Find missing baselines & datasets |
   | `qa` | `/5-osp-qa` | Multi-aspect Q&A (loops over criteria) |

Report a problem