# Overview Context engineering is the practice of managing the total state with which AI operates. This includes prompts (see: [[Prompt Engineering]]), retrieved data, [[Model Context Protocol (MCP)]] access, memory, environmental signals, etc. Some areas of context engineering include: - *Context Setup* - curating context for AI, including system prompts, [[Few-shot Classification]] examples, and available tools. - *Context Management for Long-Horizon Tasks* - addresses issues around managing model context limits by employing techniques, such as context summarization, structured note-taking, and sub-agent architectures. - *Dynamic Information Retrieval* - approaches for just-in-time context retrieval when the AI autonomously loads external data when relevant. # Key Considerations # Implementation Details # Useful Links # Related Topics ## Reference - [advanced-context-engineering-for-coding-agents/ace-fca.md at main · humanlayer/advanced-context-engineering-for-coding-agents · GitHub](https://github.com/humanlayer/advanced-context-engineering-for-coding-agents/blob/main/ace-fca.md) #### Working Notes #### Sources