EvoMaster
Build self-evolving agents for scientific research and discovery
All Skills in This Repository (4)
Pure Emerald Level IndicatorsExtract text and tables from PDFs using Python libraries.
rag
Load FAISS indexes and retrieve content via dot-path queries.
skill-creator
Create, organize, and package modular Claude skills with Python scripts.
mcp-builder
Bootstrap standardized MCP servers for Python FastMCP and Node/TypeScript MCP SDK.
Frequently Asked Questions
FAQPage SchemaHow to install EvoMaster?โผ
Run `npx skills add sjtu-sai-agents/EvoMaster --all -g -y` in your terminal to install all skills in this suite globally.
What is EvoMaster used for?โผ
EvoMaster is a framework for building autonomous research agents that can plan tasks, execute tools, and self-improve over repeated runs. It powers scientific agents for machine learning, web research, and materials science.
Can EvoMaster agents improve themselves automatically?โผ
Yes. EvoMaster supports run-level self-evolution, where agents analyze their own execution history, generate improved prompts and skills, and rerun tasks with the upgraded configuration.
Does EvoMaster support MCP tools and external APIs?โผ
Yes. Agents can connect to any MCP server through a simple config file, and the suite includes a dedicated skill for building high-quality MCP servers in Python or TypeScript.
Do I need deep coding experience to use EvoMaster?โผ
Basic Python knowledge helps, but the framework is designed for rapid setup with plug-and-play components. You can launch a working custom agent with roughly 100 lines of configuration and code.
Related Repositories in Education & Research
View All in Education & Researchโlearn-claude-code
Learn to build AI agent harnesses from scratch
ai-agents-for-beginners
Learn to build, deploy, and secure AI agents step by step
33-js-concepts
Master 33 essential JavaScript concepts with tested, verified guides