EvoSkills skill
A set of 16 skills for research work: ideation, experiments, paper writing, review, rebuttal, slides and research memory.
- 488 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- The skills run experiments and code, write project files and call literature databases.
- Some skills use outside APIs such as Gemini for illustrations.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/evoskillsDetects 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.
- The skills run experiments and code, write project files and call literature databases.
- Some skills use outside APIs such as Gemini for illustrations.
This is third-party code. Review the repository files before installing.
What it does
EvoSkills is the official skill repository for the EvoScientist agent. Each skill is self-contained: `research-ideation` finds ideas from literature and ranks them with a tournament, `paper-planning` builds a paper outline, `experiment-pipeline` runs an experiment in four stages, `experiment-craft` debugs failures and `experiment-iterative-coder` refines code in a plan, code, evaluate, refine loop. For text there are `paper-writing`, `paper-review`, `paper-rebuttal`, `paper-figures` with matplotlib charts and `academic-slides`. For reading and literature review there are `paper-navigator`, `research-survey` and `paper-graph`, which draws a lineage of a field as a Mermaid diagram. `evo-memory` keeps research memory across cycles, `nano-banana` makes slides and illustrations with a Gemini model and `evomath-tao` runs olympiad-grade proofs. Install: `npx skills add EvoScientist/EvoSkills` puts the skills into Claude Code, OpenCode, Cursor, Codex and Gemini CLI. The `mcp/` directory holds server definitions for arXiv, Exa, Perplexity, Context7 and others.
Who it is for. Researchers and graduate students who prepare machine learning papers and run experiments together with Claude Code, Codex or Cursor.
Good fit when
- An idea must be found from the literature and the best one picked through ranking.
- Experiments should run in stages: plan, run, failure analysis, code iterations.
- A paper must be split into sections, charts drawn from tables and slides prepared.
- After reviews a response draft is needed, and a paper draft needs checking before submission.
Not a fit when
- One skill is enough and a set of 16 is wider than needed.
- The work has nothing to do with science and papers.
- There is no Gemini key for illustrations with the `nano-banana` skill.
Example request
Use research-ideation to find three ideas for faster inference, rank them with a tournament and build a paper outline for the best one.Limitations
The skills target the EvoScientist agent, where they reinforce each other and `evo-memory` evolves across research cycles. In other agents the skills work individually. `nano-banana` creates illustrations through Gemini, which needs a key. The skills work with literature databases and MCP servers such as arXiv, Exa and Perplexity that have their own rules and keys. For responses to reviews, check the venue policy on generative assistance. The last commit in the repository is from October 2026.
How to disable. Delete the needed skill folders from the agent skills directory: the full set lives in `skills/`.
FAQ
How do I install EvoSkills in Claude Code?
With `npx skills add EvoScientist/EvoSkills`: skills.sh installs the skills into Claude Code, OpenCode, Cursor, Codex, Gemini CLI and other agents.
Is EvoScientist required?
No, the skills are compatible with any agent that reads SKILL.md. Inside EvoScientist they work better because `evo-memory` evolves across cycles.
How do I install one skill?
In EvoScientist with `/install-skill EvoScientist/EvoSkills@skills/paper-planning` and for other agents with the skill selection flag of `npx skills add`.
What does paper-graph do?
It builds the lineage of a research field as a Mermaid diagram showing which works build on which.
Which MCP servers are described in the repository?
The `mcp/` directory holds configurations for arXiv, alphaXiv, Exa, Perplexity, Context7, DeepWiki, LangChain docs, Sequential Thinking and SSH.
Discussion
No questions or reviews yet.
Related
Skillscafe3310's personal skill collection: spec-driven project work, reviews, reports, browser, 3D and video in one repository
Five skills from one author: arXiv reports, technical diagrams, a report site, a design system library and agent guardrails
Seven skills for research, HTML explainers and source citation, installable all at once or one at a time through npx.
A collection of skills for decisions and projects: strategy, market research, thinking, data analysis, docs and slides.
293 skills in 16 categories for Codex, Claude Code and Cursor with an npx installer, provenance and license checks
Third Brain Skills: a knowledge base and verification for agents
Third Brain Skills
21 skills build an Obsidian-style knowledge base with provenance, daily loops and verification before any done claim
README in short
The full research cycle in one skill set: from ideas and experiments to the paper, review and slides, plus MCP server configurations.