Context Management
Engineering what the model sees at inference time.
Context management is the work of deciding what information goes into a model’s context window at inference time — context engineering, windowing, and guarding against failure modes like context poisoning and context rot that degrade an agent’s reasoning.
Context management includes system instructions, tool definitions, retrieved data and conversation history. Anthropic's context-engineering guide describes curating these together and loading information as needed, with compaction and external notes for longer tasks (Anthropic, Effective context engineering for AI agents). A practical design exercise is to list what must survive a task handoff: the goal, constraints, evidence and unresolved decisions. Keep that list alongside your tests for context changes.
Michel Tricot in episode 52 demonstrates the costs of giving an agent poorly shaped data through raw API calls. Kapil Chhabra in episode 75 distinguishes data definitions from the business context needed to interpret them. What is context in AI agents? gives the foundation; prompt vs context vs memory engineering separates the engineering responsibilities. For information that persists between tasks, continue to agent memory.
Try it in your own project: repair a stale handoff, copy the template, and check the next session.
7 episodes
- Why Context Alone Isn't Enough for Enterprise AI Agents | WisdomAI CPO Kapil Chhabra
- Slack Wants to Be the Context Harness for Code | CPO Jaime DeLanghe
- The AI Framework Era Is Over: Why Context Is the Moat | Jerry Liu of LlamaIndex
- Agent Memory: The Last Battleground in the AI Stack | Richmond Alake, Oracle
- Context Poisoning is Killing Your AI Agents: How to Stop it | Michel Tricot, Airbyte
- Architecting AI Agents: The Shift from Models to Systems | Aishwarya Srinivasan
- From Demo to Defensibility: How to Build an AI Business that Lasts | Aurimas Griciūnas
Explainers on this topic
- What a Context Window Is, and Why It Fills Up
- How to Make Agent Memory Portable Across Apps
- How to Brief an AI Assistant
- How to Prepare PDFs for RAG
- Semantic Layer vs. Context Layer for AI
- What Is Data Lineage for AI?
- Can an AI Agent Run Your Workday?
- Context Poisoning: Why Good Models Still Fail
- Do You Still Need an AI Agent Framework?
- Prompt vs. Context vs. Memory Engineering
- The 5 Sources of Context an AI Agent Needs
Terms on this topic
- Agent Instruction File
- Context Compaction
- Context Engineering
- Context Layer
- Context Poisoning
- Context Window
- Data Catalog
- Entity Resolution
- Long Context
- Lost in the Middle
- Multi-Turn Conversation
- Prompt Caching
- System Prompt
Essays on this topic
Guests on this topic
Kapil ChhabraJaime DeLangheJerry LiuRichmond AlakeMichel TricotAishwarya SrinivasanAurimas Griciūnas