Co-Researcher plugin
Academic plugin for Claude Code, Gemini CLI, Codex and OpenCode: literature reviews, peer review and citation checks
- 138 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Runs a verification script and writes review files into the project working folder
- Calls OpenAlex, arXiv, Europe PMC and Crossref over the network
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/co-researcherDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Runs a verification script and writes review files into the project working folder
- Calls OpenAlex, arXiv, Europe PMC and Crossref over the network
This is third-party code. Review the repository files before installing.
What it does
Co-Researcher runs academic reviews and peer review inside the agent you already use. Searches run against real scholarly databases OpenAlex, arXiv and Europe PMC, and every bibliography passes a verification gate that catches fabricated, mismatched and retracted citations before output. The /research, /analyze and /review commands (in Codex $research, $analyze, $review) start a research project, a critical analysis and a peer review. /research first drafts a plan and waits for your approval: --auto skips it, --plan-only stops at the plan and --template=quick|rigorous|comprehensive sets depth. Everything else starts by describing the task in plain words and the matching one of fifteen skills starts on its own: literature review, PRISMA systematic review, critical analysis, peer review, methodology, hypotheses, quantitative and qualitative analysis, ethics, grants and academic writing. For a literature review the agent writes a protocol, a corpus of papers, full texts, a synthesis, claims with verbatim quotes and a bibliography into a review/{slug}/ workspace. Before delivery validate_review.py checks without a model that every quote occurs in its paper and that numbers and direction in the synthesis agree with a verified claim. The result arrives as valid, valid with warnings or INVALID.
Who it is for. Graduate students, researchers and editors who need a literature review or peer review with checkable references.
Good fit when
- You need a literature review on a topic with a search protocol, paper screening and synthesis
- A manuscript has to be peer reviewed for methodology and argument logic
- Before submitting text you want to be sure the bibliography has no fabricated or retracted works
- A systematic review under the PRISMA standard is being prepared with a protocol and inclusion criteria
- A grant proposal needs research aims, hypotheses and falsification criteria
Not a fit when
- You need a quick answer to a casual question: the full pipeline with verification takes time
- Full paper texts are unavailable: the agent reads only what it can retrieve through legal open-access routes
- You want text with no source checking: the skill always leads to citation verification
- You work in an agent absent from the install table: the suite targets four platforms
Example request
/research "remote work and productivity" --template=rigorous and show me the plan for approvalLimitations
- Search is limited to OpenAlex, arXiv and Europe PMC and verification to OpenAlex, Europe PMC and Crossref, so work outside those databases and paywalled papers may end up supported by an abstract only.
- If a check fails the agent receives the findings and up to three rounds to repair the review, and the result can still arrive with warnings or as INVALID.
- Verification needs Python and network access; on Codex and OpenCode installation follows the separate .codex/INSTALL.md and .opencode/INSTALL.md instructions.
- With npx skills add for other agents only the skills are installed, without commands and the session-start hook.
How to disable. In Claude Code run /plugin uninstall co-researcher, in Gemini CLI remove the extension with gemini extension uninstall co-researcher. The review/{slug}/ folders in your projects stay and are deleted manually.
FAQ
How do I install Co-Researcher in Claude Code?
Run /plugin marketplace add poemswe/co-researcher, then /plugin install co-researcher. The /research, /analyze and /review commands appear right away.
Which agents are supported?
Claude Code, Gemini CLI, OpenAI Codex and OpenCode. In Gemini CLI install with gemini extension install https://github.com/poemswe/co-researcher; Codex and OpenCode have instructions at the README links.
How does the plugin catch fabricated citations?
validate_review.py checks without a model that every quote occurs verbatim in the paper and every citation points to exactly one screened paper. The bibliography is matched against OpenAlex, Europe PMC and Crossref.
Can I skip waiting for plan approval?
Yes, add the --auto flag to /research. The --plan-only flag stops at the plan and --template=quick|rigorous|comprehensive sets depth.
What does valid with warnings mean?
All checks passed but caveats are listed beside the review, such as support from an abstract only or a bibliography that could not be resolved online.
Discussion
No questions or reviews yet.
Related
PluginsFeynman — a terminal AI agent for scientific research: literature reviews, source checks, paper critique and comparing a paper with its code
A research system on top of Codex CLI or Claude Code: runs of experiments, discussion of progress and a paper draft.
A Claude Code plugin for autonomous research: from a topic to ideas, proposals and running experiments in closed agent loops.
CLI agent for literature reviews: searches 11 open databases, screens papers and writes a bilingual cited report
A set of 11 Claude Code skills that audits a third-party paper for self-consistency and fabrication and writes a report.
Claude Code supervises while Codex, opencode and pi work on your quota: delegated edits, independent review, model discussions
README in short
The page presents the suite for four agents, gives an install table and the /research, /analyze and /review commands, then lists fifteen skills with the Systemic Honesty principles. The main part covers the literature review with checks: working files in review/{slug}/, the validate_review.py rules and the three result statuses.