Academic research plugin with 15 skills; literature reviews check citations and quotes against OpenAlex, arXiv and Europe PMC.
The literature-review skill is the real substance, with scripted verification and a large test set; the rest of the pack is lighter prompt material.
Co-Researcher is a plugin for Claude Code, Gemini CLI, Codex and OpenCode that aims at research output you can check. The core is the literature-review skill. The agent searches OpenAlex, arXiv and Europe PMC, screens candidates, pulls full text through open-access routes, and writes a workspace under review/{slug}/ with a protocol, a screened corpus, per-paper full text, the synthesis, a claims file with the verbatim quote behind every cited sentence, and a checked bibliography. Before delivery it runs validate_review.py, a non-model check that every quote occurs in its paper and that numbers, negation and direction in a cited sentence match the claim. Failed checks go back to the agent for up to three repair rounds.
The other 14 skills are much shorter prompts (roughly 2-11 KB each) for systematic review (PRISMA), peer review, critical analysis, hypothesis testing, quantitative and qualitative methods, ethics review, grant writing, research-synthesis, plus an academic-writing skill that strips AI-sounding prose and a research-manager for resumable project state. Slash commands are /research, /analyze and /review; a session-start hook loads the shared honesty principles. The literature scripts run through uv and can import a Zotero collection, BibTeX or a PDF folder.
One 30 KB literature-review skill with 10 reference files and 21 scripts including a validator, plus 577 eval files in the repo, while the other 14 skills are 2-11 KB prompts without scripts.
Counted from the repository file tree. How we score depth.
Academic research skills; the literature-review skill runs local Python scripts that query public scholarly APIs (OpenAlex, arXiv, Europe PMC, Crossref, DataCite) with your search terms and identifiers, and can read a local Zotero library. It needs uv (installed via a setup script or Astral's official installer) and an optional OpenAlex key that is sent only to OpenAlex. Most of the repo was not available to us.
curl -LsSf https://astral.sh/uv/install.sh | sh (Astral's official installer).skills/literature-review/SKILL.md:30The repository (703 files) on Oct 29, 2026. Not a full review: 655 of 703 files were not downloaded (large repository; we reviewed the skill folders); we read 11 of 21 code files; it relies on code outside the repository: PyPI packages the scripts import (bibtexparser, pymupdf, python-dotenv and others) installed through uv; scripts/setup.sh and the other plugin-level files: 655 readable files in the repo were not available to us; Parts of the larger scripts (check_claims.py 60 KB, review_integrity/ modules, validate_review.py) were searched for network, process and file-write calls rather than read line by line.
NVIDIA SkillSpector static scan: 25 pattern hits (Data Exfiltration 12, Dangerous Code Execution 5, Excessive Agency 3, Privilege Escalation 2, Data Flow 1, MCP Rug Pull 1). The CRITICAL tainted-flow hit is http_client.py's urllib request to the scholarly API host (with an optional OpenAlex key); 'Env Variable Harvesting' is validate_review.py copying the environment to its git subprocess; credential hits are the optional OPENALEX_API_KEY and ~/.env loading; the astral.sh hit is the documented uv installer fallback.
A review, not a certification. Skills change after we check them — read the files before installing.
MIT — commercial use allowed.
Not really. The /research command drafts a plan and waits for approval unless you pass --auto, and the validator retry loop is capped at three rounds, so it suits supervised runs more than unattended ones.
Five very detailed Claude Code skills for the academic pipeline: deep research, paper writing, peer review, revision and systematic-review screening.
K-Dense's library of 177 science skills covering bioinformatics, chemistry, clinical research, data analysis, lab tooling and scientific writing.
A single research skill with an 8-phase pipeline, parallel search, source scoring, claim and citation checks, and report output in Markdown, HTML and PDF.
Co-Researcher is an open-source academic research plugin with 15 skills for literature reviews, systematic reviews, peer review, hypothesis testing and grant writing. Its literature-review skill writes a workspace with claims, quotes and a bibliography, then runs a validator against it.
In Claude Code run /plugin marketplace add poemswe/co-researcher and then /plugin install co-researcher. Gemini CLI uses gemini extension install with the GitHub URL, Codex and OpenCode use the INSTALL.md files in the repo, and npx skills add poemswe/co-researcher installs the skills alone.
It checks bibliographies against OpenAlex, Europe PMC and Crossref and verifies that each quoted passage occurs in the retrieved paper text. That removes many fabricated or mismatched references, but the README itself says abstract-only support and unresolved references are reported as warnings, so you still need to read the flagged items.
Co-Researcher puts most of its weight on one verified literature-review pipeline with scripts and tests. Imbad0202/academic-research-skills is organised as a research, write, review, revise and finalize flow. Pick by whether you need citation verification or a fuller paper-writing cycle.