Research Agent Skills
A library of 280+ research skills: literature search, experiment design, analysis, reference checks and paper preparation
- 124 ★ on the repository
- Last commit
Extended access
Why this level
- Skills run scripts and call the external Crossref, OpenAlex and arXiv databases
- Third-party software and publisher templates in some skills carry their own licenses
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/research-agent-skillsDetects 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.
- Skills run scripts and call the external Crossref, OpenAlex and arXiv databases
- Third-party software and publisher templates in some skills carry their own licenses
This is third-party code. Review the repository files before installing.
What it does
Research Agent Skills gives an agent the methods of a careful scientist. Each skill is a folder with a `SKILL.md`, reference material and tested scripts, and the agent opens it only when the task fits. Together they cover the path from question to publication: hypotheses, literature reviews and PRISMA, experiment design, data analysis in biology, chemistry, medicine, physics and machine learning, writing, journal preparation and peer review. Three skills stand out. `citation-verification` checks every reference against Crossref, OpenAlex and arXiv, `unslop-academic-writing` removes stock phrasing and empty emphasis from academic prose, and venue skills for Nature, Cell, Science, IEEE, ACM, Springer, Elsevier, APA, PLOS and arXiv make the agent reread current guidelines before formatting. Every skill is scanned for prompt injection and network calls. It installs with one skills.sh command, or as a Claude Code plugin marketplace with bundles by field.
Who it is for. Researchers, graduate students and scientific editors who give an agent literature, statistics and paper formatting and want verifiable methods.
Good fit when
- `refs.bib` holds hundreds of references and you need to catch fabricated or corrupted ones before journal submission
- You are about to plan an experiment: size the sample and pre-register hypotheses
- A paper is formatted for Nature or IEEE and the agent must first read the current guidelines
- An introduction reads as boilerplate and must be rewritten around evidence without changing the claims
Not a fit when
- You want the agent to run the experiment itself or replace the researcher's judgment: the set supports method and does not replace a person
- You need one small skill: the library is large, so install only the skills you need with `--skill`
- You need clinical decisions: the `treatment-plans` skill only formats decisions a clinician already made
Example request
Check every reference in refs.bib against Crossref and arXiv and list the verified, mismatched and not-found ones.Limitations
The author calls part of the model-based results pilots on small samples, so style quality and trigger accuracy deserve a recheck on your own texts. The skills do not run experiments or replace a researcher's judgment. Installing through skills.sh needs Node.js 18 or newer, and through Python it needs `uv` or `pip`. References are checked against the external Crossref, OpenAlex and arXiv databases, and some skills document third-party software or bundle publisher templates with their own licenses. The project license is MIT, which GitHub does not detect automatically.
How to disable. Remove the installed skills from the agent's skills directory or the plugin through /plugin. Delete scripts and data the skills left in the project by hand.
FAQ
How do I install only the skills I need?
Run `npx skills add KalarisLabs/research-agent-skills --skill citation-verification apa7`. The `--list` flag shows the skills without installing, and `--agent claude-code` limits it to one agent.
How do I install it through Python?
Run `uvx research-agent-skills install citation-verification`, or `pip install research-agent-skills` and then `research-agent-skills install --bundle research-essentials`.
Are there field bundles in Claude Code?
Yes. The repository works as a plugin marketplace: `/plugin marketplace add KalarisLabs/research-agent-skills`, then for example `/plugin install biology-research@research-agent-skills`.
What does citation-verification check?
Every reference against Crossref, OpenAlex and arXiv, including preprints without a DOI. On the author's own labelled benchmark it flagged every fabricated and corrupted reference and passed every correct one.
Do 280 skills slow the agent down?
No. The agent sees one short description per skill and opens the full text only when the task fits, so unused skills cost almost nothing.
Discussion
No questions or reviews yet.
Related
SkillsA skill for master and PhD students that runs research through novelty and reality gates, from review to claim checks
A marketplace of five Claude Code plugins for researchers: literature search, research design, project memory and manuscript work
Seven skills for planning research, reviewing literature, preparing experiment code and drafting a paper with traceable sources
A Claude Code skill for scientific literature reviews: an interview, Scopus and OpenAlex search, DOI checks and a final docx file
Four skills for research text: drafting a paper from data, English polishing, removing templated phrasing and slides from a paper
Research Skills gives Claude Code, Codex and Copilot CLI plugins for research: literature, grants, papers, figures, slides.
README in short
The README opens with a one-command install and the problems the skills solve: fabricated references, boilerplate text, the wrong method, journal rules, unreproducible analysis and unsafe third-party skills. It then walks through the research stages, a full study of a fruit fly brain, a catalog of skills by field and common questions.