AI Scientist

Build AI agents that automate scientific research using [AIScientistv2](https://github.com/SakanaAI/AIScientistv2) — an agentic tree search framework for hypothesis generation, experiment design, data analysis, and paper writing.

概览

The AI Scientist skill, available through the TerminalSkills/skills repository, enables developers to build agents capable of automating the end-to-end scientific research process. It leverages the AIScientistv2 framework, which utilizes an agentic tree search approach to systematically handle complex tasks. These tasks include generating novel hypotheses, designing rigorous experiments, performing data analysis, and drafting full scientific papers. By integrating this skill, compatible agents like Claude, Gemini, and Cursor can navigate the research lifecycle autonomously. The implementation focuses on structured reasoning and iterative refinement, providing a programmatic way to scale scientific discovery. This resource is part of a collection of modular skills hosted on GitHub, where the parent repository currently maintains a popularity of 71 stars.

使用场景

Automating the generation and refinement of scientific hypotheses using agentic tree search.
Streamlining the end-to-end process of experiment design and data analysis for research papers.
Drafting and reviewing comprehensive scientific manuscripts based on experimental results.

安装说明

# Review source first
open https://github.com/TerminalSkills/skills/blob/main/skills/ai-scientist/SKILL.md

Copy or clone the skill folder into your agent skills directory after reviewing its instructions and scripts.

安全提示

Users should review the AIScientistv2 framework's execution environment requirements, as automated experiment design and data analysis may involve running code or accessing external APIs. Ensure that the agent's permissions are restricted to necessary directories and that any generated scripts are audited before execution in sensitive environments.

相关 Skills

Skill Creator

anthropics/skills

编程开发

创建新技能、修改和改进现有技能,并衡量技能性能。适用于用户希望从零开始创建技能、编辑或优化现有技能、运行评估以测试技能、通过方差分析对技能性能进行基准测试,或优化技能描述以提高触发准确性的场景。

Claude CodeClaude
pythondesign
165,880 StarsApache-2.0

Claude API

anthropics/skills

编程开发

Claude API / Anthropic SDK 参考指南 — 包含模型 ID、定价、参数、流式传输、工具使用、MCP、智能体、缓存、Token 计数、模型迁移。 TRIGGER — 阅读 BEFORE 并打开目标文件;不要因为其“看起来像单行代码”而跳过 — 适用场景:提示词以任何形式提及 Claude/Anthropic(Claude、Anthropic、Fable、Opus、Sonnet、Haiku、`anthropic`、`@anthropic-ai`、`claude-*`、`us.anthropic.*`、`[1m]`);用户询问关于 LLM(定价/模型选择/限制/缓存) — 绝不回答

Claude CodeClaude
typescriptpython
165,880 StarsApache-2.0

MCP Builder

anthropics/skills

编程开发

创建高质量 MCP (Model Context Protocol) 服务器的指南,使 LLM 能够通过设计良好的工具与外部服务进行交互。在构建 MCP 服务器以集成外部 API 或服务时使用,无论是使用 Python (FastMCP) 还是 Node/TypeScript (MCP SDK)。

CodexClaude
typescriptpython
165,880 StarsApache-2.0

CI Cd And Automation

addyosmani/agent-skills

DevOps

自动化 CI/CD 流水线设置。适用于设置或修改构建和部署流水线。当需要自动化质量门禁、在 CI 中配置测试运行器或建立部署策略时使用。

typescriptsecurity
81,373 StarsMIT

API And Interface Design

addyosmani/agent-skills

编程开发

指导稳定的 API 和接口设计。在设计 API、模块边界或任何公共接口时使用。在创建 REST 或 GraphQL 端点、定义模块间的类型契约或建立前后端边界时使用。

typescriptfrontend
81,373 StarsMIT

Browser Testing With Devtools

addyosmani/agent-skills

编程开发

通过 Chrome DevTools MCP 在真实浏览器中进行测试。在构建或调试任何在浏览器中运行的内容时使用。当你需要检查 DOM、捕获控制台错误、分析网络请求、分析性能或使用真实运行时数据验证视觉输出时使用。需要配置 chrome-devtools MCP 服务器。

Claude CodeClaude
javascriptreact
81,373 StarsMIT