research-agent-skills

Turn any AI agent into a research co-scientist: 280+ open-source agent skills for literature reviews, hypotheses, experimental design, data analysis, paper writ

What is research-agent-skills?

Turn any AI agent into a research co-scientist: 280+ open-source agent skills for literature reviews, hypotheses, experimental design, data analysis, paper writing, citation verification and peer review. Claude Code, Codex, Cursor, Gemini CLI and any Agent Skills harness. npx skills add KalarisLabs/research-agent-skills

Overview of research-agent-skills

What it is

Research Agent Skills is a free, open-source library of over 280 agent skills maintained by Kalaris Labs. It is designed to turn any AI agent (such as Claude Code, Codex, Cursor, Gemini CLI, or Copilot) into a research co-scientist. Each skill is a folder containing a SKILL.md playbook, reference material, and tested scripts. The agent loads a skill only when the task requires it, allowing the same agent to perform literature reviews, design experiments, analyze data, write papers, verify citations, and conduct peer review.

Key features

  • 280+ open-source agent skills covering literature reviews, hypothesis generation, experimental design, data analysis, paper writing, citation verification, and peer review.
  • Citation verification skill checks every reference against Crossref, arXiv, and DOI-less preprints before inclusion in a draft.
  • Unslop academic writing skill removes patterns that reviewers recognize as AI-generated text, reducing the slop index by 83% in a pilot study.
  • Venue skills for journal formatting and submission rules, including Nature, Cell, Science, IEEE, ACM, Springer, Elsevier, APA, PLOS, and arXiv.
  • Every skill is scanned for prompt injection, payloads, and unreviewed network calls using Cisco skill-scanner, CodeQL, and an injection linter on every pull request.
  • Skills carry tested scripts, reproducibility statements, and explicit verification steps; each skill passes a static quality gate with a mean score of 95.2/100.

Use cases

Researchers can use these skills to generate and stress-test hypotheses, search papers and databases, run PRISMA systematic reviews, plan experiments with statistical power analysis, analyze data using field-standard tools (e.g., scanpy, rdkit, astropy, pydicom, pytorch-lightning), draft and visualize scientific writing, verify every reference and reproducibility claim, format submissions for specific venues, prepare arXiv submissions, write grants, review papers, and answer reviewers. The skills are built to address three common failures of AI research help: AI slop, hallucinated citations, and outdated venue rules.

Getting started

bashnpx skills add KalarisLabs/research-agent-skillsOne command via skills.sh, no account required. You can pick all skills or only the ones you need. Alternatively, paste the following into your agent:textUse the skills in 'https://github.com/kalarislabs/research-agent-skills' that are relevant to the current task. Run `npx skills add 'https://github.com/kalarislabs/research-agent-skills'` and select the relevant skills, then follow their instructions.

License and pricing

The license is not stated in the provided metadata or README. The pricing type is listed as open_source.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/kalarislabsresearch-agent-skills.svg)](https://yunodb.com/tools/kalarislabsresearch-agent-skills?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/kalarislabsresearch-agent-skills?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/kalarislabsresearch-agent-skills.svg" alt="Featured on YunoDB" height="20"></a>