Meirtz/Awesome-Context-Engineering logo

Meirtz/Awesome-Context-Engineering

Free

curated papers, frameworks, and implementation guides

FreeFree tier
Type
Open Source

About Meirtz/Awesome-Context-Engineering

A comprehensive survey and collection of resources on Context Engineering, tracing the evolution from static prompt engineering to dynamic, context-aware AI systems and further into agent runtimes, memory systems, protocols, coding agents, and observability stacks. The repository includes hundreds of papers, frameworks, and implementation guides for LLMs and AI agents, organized into sections on long context, RAG, memory, agent communication, tool use, evaluation, and applications. It also covers production-grade topics such as agent harnesses, runtime systems, context management, memory artifacts, and open protocols like MCP and A2A.

Key Features

Curated collection of hundreds of papers, frameworks, and implementation guides
Covers long context, RAG, memory, agent communication, tool use, evaluation, and applications
Includes agent era update with harnesses, runtime systems, context management, and memory artifacts
Discusses open protocols such as MCP, A2A, AG-UI, and ACP
Published survey paper on arXiv and Hugging Face Papers
Community support via WeChat group and Discord server

Pros & Cons

Pros
  • Comprehensive coverage of context engineering theory and practice
  • Regularly updated with new paradigms like agent engineering
  • Includes an academic survey paper for deep understanding
  • Active community with discussion channels
  • Free and open-source resource
Cons
  • Primarily a resource collection; not a tool or platform itself
  • Requires technical background to fully leverage the materials
  • Some links and references may become outdated over time

Best For

Research and study of context engineering techniques and evolutionBuilding production-grade AI systems with dynamic context managementDeveloping AI agents with runtime state, memory, and tool integrationUnderstanding modern prompt engineering and its limitationsImplementing observability and evaluation for LLM-based applications

FAQ

What is Context Engineering?
Context Engineering is the evolution from static prompt engineering to dynamic, context-aware AI systems. It encompasses the complete information payload provided to LLMs at inference time, including all structured components necessary for plausible task completion.
Does this repository include an academic paper?
Yes, a survey paper titled 'A Survey of Context Engineering for Large Language Models' has been published and is available on arXiv and Hugging Face Papers.
How can I get involved or ask questions?
You can join the WeChat group or Discord server for discussions, or open an issue on the GitHub repository. Contact the author via email or through the repository.