context window composition

4 articles · 15 co-occurring · 0 contradictions · 0 briefs

Article explicitly defines what comprises the context window: system prompt, conversation history, tool calls/outputs, and file reads. This is a foundational concept for understanding context engineer

Related concepts

context window management 4 prompt engineering 3 retrieval augmented generation 2 multi agent orchestration 2 agent architecture 2 tool use and function calling 1 tool integration patterns 1 state management 1 session lifecycle management 1 rag retrieval augmented generation 1 rag pattern 1 memory persistence 1 information persistence across sessions 1 context rot 1 context preservation tradeoff 1

Evidence chain (4 articles, showing 4)

Day 1 — Welcome to the AI Era: The 2026 Landscape | by Neeraj Kushwaha | May, 2026 | Medium example_of

Article explicitly defines context window as sum of system prompt + conversation history + retrieved docs + user input, which is a direct operationalization of context composition

Effective Context Engineering for AI Agents: The Missing Layer in AI Systems | Fluid AI Blog example_of

Article explicitly describes what should be in context windows: system instructions, user queries, retrieved knowledge, conversation history, tool outputs, structured data, memory records.

@dani_avila7: An article you need to read if you're serious about managing context in your ... example_of

🦸🏻#14: What Is MCP, and Why Is Everyone – Suddenly!– Talking About It? example_of

MCP is a practical pattern for composing multiple external context sources (Drive, GitHub, Slack, Postgres) into a single AI context window.

query this concept

$ db.articles("context-window-composition")

$ db.cooccurrence("context-window-composition")

$ db.contradictions("context-window-composition")