Scientific Agents plugin
Over 500 AGENTS.md profiles, each teaching an AI agent to work like a senior practitioner in one scientific profession
- 245 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- Profiles contain text instructions only and run nothing
- The instructions shape the agent's conclusions in science, so check facts and calculations
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add plugins/scientific-agentsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- Profiles contain text instructions only and run nothing
- The instructions shape the agent's conclusions in science, so check facts and calculations
This is third-party code. Review the repository files before installing.
What it does
Scientific Agents is a collection of K-Dense AGENTS.md files: one per profession, from astrophysicist and acoustician to epidemiologist. A profile describes how a practitioner frames a problem, which databases, instruments and methods they reach for, how they stress-test claims, hunt for failures and report results. It is a working procedure for the agent, not an abstract for a reader. Each profile sits in its own folder packaged as a plugin with a skill, so in Claude Code it installs from the marketplace, for example /plugin install astronomer@scientific-agents. In other tools (Cursor, Codex, OpenCode, Pi and Agent Plugins-compatible ones such as Copilot and VS Code) the file is copied or linked as AGENTS.md at the project root. The agent reads it at the start of every session. Profiles are built from per-profession research instead of generic calls for rigor. The authors published an evaluation on scientific tasks as an arXiv paper.
Who it is for. Researchers, engineers and analysts who do scientific or engineering work with an AI agent and want it to know the methods, sources and reporting norms of their field.
Good fit when
- A project analyses sequencing data and the agent needs a bioinformatician's context: controls, statistics, sources
- A methods section is being drafted and the text should sound like a practitioner of the field
- An engineer debugs an experiment and wants the agent to hunt for failures the way a specialist does
- A lab group wants one shared AGENTS.md that holds its reporting norms
Not a fit when
- You need a tool that runs calculations or analysis: a profile shapes how the agent thinks and computes nothing itself
- The task lies outside science and engineering: the repository may have no profession for it
- You need medical or legal advice: a profile does not replace an expert
Example request
Install the clinical epidemiologist profile and use its method to set up the baseline data table for the studyLimitations
A profile only adds context to the agent prompt. It does not connect the databases, instruments and computing packages it mentions, and answer accuracy depends on the model. The authors tested the profiles on scientific tasks in their paper, so on your own field you will have to judge the result yourself. Each profession needs its own profile, and combining two means merging their rules by hand in one AGENTS.md. The last commit in the repository is from October 2026.
How to disable. Delete the profile AGENTS.md from the project root or run /plugin uninstall for the profession plugin in Claude Code.
FAQ
How do I pick a profile?
Open the Agents table in the README and find your profession, for example developmental-biologist or clinical-epidemiologist. Then install it as a plugin or copy its AGENTS.md into the project.
What is AGENTS.md?
A plain Markdown file of instructions for an AI agent in a project. The format is open and works across many tools, so you do not need a separate prompt for each product.
Can a profile be combined with project rules?
Yes. Shared lab or repository conventions go in the same AGENTS.md or a separate file the tool can include, as the README explains per tool.
How do I update installed profiles?
In Claude Code run /plugin marketplace update scientific-agents.
What are the profiles based on?
Research per profession: theory, ways of framing problems, tools and instruments, data sources, statistics, known failure modes and reporting standards. The authors published an evaluation on scientific tasks on arXiv.
Discussion
No questions or reviews yet.
Related
PluginsA multi-agent harness that verifies results by artifacts: stop hooks, an independent judge and an event log for Claude Code, Codex, Cursor and more
A base AGENTS.md and 12 skills: delegation to Codex, Grok, Devin and Antigravity, planning, copywriting, scoped changes.
claude-code-sub-agents
Claude Code Subagents Collection
A collection of specialized subagents for Claude Code: frontend, backend, data, infrastructure, testing, security
A scaffold of personas, rules and skills in plain Markdown: Maestro hands work to different models and spares your main plan
A skill that distills chats, documents and public sources into a reusable person profile for AI agents and compatible bots
An open coding agent for the terminal, browser and desktop that works with models from OpenAI and Claude to Ollama
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Scientific Agents (this card) | no | 245 | |
| Scientific Agent Skills | no | 56,977 | skills for scientific computing and tools; Scientific Agents sets a professional way of thinking |
| Claude Scientific Writer | no | 2,571 | helps write scientific text; profiles are broader and cover the whole workflow |
README in short
The README explains what AGENTS.md is, how profiles are built (scoping, research, writing) and how to use one in three steps. It then gives instructions for Agent Plugins, Cursor, Claude Code, Antigravity, Codex, OpenCode and Pi, and a table of professions.