What is this?
This repository holds a write-up from HumanLayer on 'context engineering' for AI coding agents: practical techniques for deciding what information to feed an agent like Claude Code so it produces better, more reliable code changes.
This repository holds a write-up from HumanLayer on 'context engineering' for AI coding agents: practical techniques for deciding what information to feed an agent like Claude Code so it produces better, more reliable code changes.Is this a tool I install, or just something to read?It is a written guide/essay rather than a software library or CLI tool, so there is nothing to install or run.
It is a written guide/essay rather than a software library or CLI tool, so there is nothing to install or run.Who is this guide meant for?Developers who already use AI coding agents and want to get better results by being more deliberate about what context they give the agent.
Developers who already use AI coding agents and want to get better results by being more deliberate about what context they give the agent.What is the research-plan-implement workflow mentioned in context engineering?It is a staged approach where the agent first researches the codebase, drafts an explicit plan, and only then implements the change, keeping each phase's context focused and reviewable.
It is a staged approach where the agent first researches the codebase, drafts an explicit plan, and only then implements the change, keeping each phase's context focused and reviewable.What does context compaction mean?It refers to periodically summarizing or trimming an agent's conversation history so old, irrelevant details do not crowd out the information that actually matters for the current task.
It refers to periodically summarizing or trimming an agent's conversation history so old, irrelevant details do not crowd out the information that actually matters for the current task.Why does context engineering matter for coding agents?Because an agent's output quality depends heavily on what is in its context window; poorly curated context can lead to confused or low-quality code changes even with a capable model.
Because an agent's output quality depends heavily on what is in its context window; poorly curated context can lead to confused or low-quality code changes even with a capable model.What are similar projects?vibe-coding-cn, harness-engineering, ai-system-design, 12-factor-agents, context-engineering-intro, claude-code-best-practice, Prompt-Engineering-Guide, awesome-harness-engineering
vibe-coding-cn, harness-engineering, ai-system-design, 12-factor-agents, context-engineering-intro, claude-code-best-practice, Prompt-Engineering-Guide, awesome-harness-engineeringWhich topics is it in?context-engineering, best-practices, coding-agents, prompt-engineering, artificial-intelligence, agentic-ai, guide, agents, documentation, markdown, tutorial, resources
context-engineering, best-practices, coding-agents, prompt-engineering, artificial-intelligence, agentic-ai, guide, agents, documentation, markdown, tutorial, resources