A browser built for AI agents and you to share: agents run tasks in isolated Spaces while your own tabs and logins stay untouched.
K-Dense-AI/scientific-agent-skills
A collection of 177 validated agent skills that teach AI coding agents to work with scientific libraries, databases and research workflows.
- Stars
- 52,771
- Forks
- 4,678
- Open issues
- 49
- Last commit
- yesterday
What is it?
Scientific Agent Skills is a library of procedural guidance files that AI agents such as Claude Code, Cursor, and Codex can load when a task calls for a specific scientific package, data source, or workflow. Each skill is a SKILL.md document, often with code examples, reference notes, helper scripts, or templates, covering areas from bioinformatics and cheminformatics to clinical research, materials science, geospatial analysis, and scientific writing.
The collection covers roughly 80 documented public and licensed databases, including PubChem, ChEMBL, UniProt, and ClinicalTrials.gov, along with over 70 Python package skills such as RDKit, Scanpy, scikit-learn, pymatgen, and GeoPandas. It follows the open Agent Skills standard and can also be installed as an Agent Plugins package. Skills are version-scoped and state their dependencies, and the project notes that the agent can still use any Python package without a dedicated skill.
Because skills can run code and change agent behavior, the project asks users to review each SKILL.md and install only the subset they need.
Key features
- 01Covers biology, chemistry, medicine, physics, engineering, Earth science and data analysis
- 02Documents about 80 public or licensed scientific databases with pagination and provenance guidance
- 03Provides version-scoped workflows for packages such as RDKit, Scanpy, PyTorch Lightning and pymatgen
- 04Follows the open Agent Skills standard and works with Cursor, Claude Code, Codex and Google Antigravity
- 05Includes a security scanning process and a security report for the skills
Use cases
- Single-cell RNA-seq analysis with Scanpy and public atlas data
- Virtual screening and molecular property prediction with RDKit and ChEMBL
- Literature review and evidence-traceable scientific writing
- Designing primers or processing mass spectrometry data in a research pipeline
Things to know
- Several clinical, healthcare and regulatory skills are research-only and are not meant for patient-specific diagnosis, treatment, or certification decisions.
- The docx, pdf, pptx and xlsx document skills are no longer bundled and must be installed separately from Anthropic's repository.
Topics
FAQ
- What is scientific-agent-skills?
- A collection of 177 validated agent skills that teach AI coding agents to work with scientific libraries, databases and research workflows. Scientific Agent Skills is a library of procedural guidance files that AI agents such as Claude Code, Cursor, and Codex can load when a task calls for a specific scientific package, data source, or workflow. Each skill is a SKILL.md document, often with code examples, reference notes, helper scripts, or templates, covering areas from bioinformatics and cheminformatics to clinical research, materials science, geospatial analysis, and scientific writing.
- Is scientific-agent-skills free to use? What license does it have?
- scientific-agent-skills is open source under the MIT license. Check the license terms before using it in commercial products.
- Is scientific-agent-skills still maintained?
- Actively maintained. The last commit to scientific-agent-skills was on Oct 19, 2026.
- What can I use scientific-agent-skills for?
- Single-cell RNA-seq analysis with Scanpy and public atlas data Virtual screening and molecular property prediction with RDKit and ChEMBL Literature review and evidence-traceable scientific writing Designing primers or processing mass spectrometry data in a research pipeline
- Where can I get scientific-agent-skills?
- The source code and installation instructions are on GitHub: https://github.com/K-Dense-AI/scientific-agent-skills
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
An agent skill that lets Claude Code or Codex produce voiceover-driven explainer videos, with word-level audio sync and Remotion rendering.
A collection of AI agent skills by Baoyu for content creation, image generation, and publishing to WeChat, X, and Weibo, for Claude Code, Codex, and similar agents.
An agent skill that turns PDFs, books, slides and papers into traceable knowledge bases and interactive learning web pages.
Open-source AI workspace for creators that combines video translation, downloading, image and video generation, voiceovers and scripting, built on Codex CLI.
An Agent Skill that gives local AI agents like Claude Code full web access via tiered tool routing, a CDP browser proxy, and per-site experience notes.