Foundations: Context Engineering and Instructions¶
The two disciplines that underpin every effective agent interaction: what information enters the context window, and how instructions shape agent behavior.
Context engineering and instructions form the foundation of productive work with AI coding agents. Getting either wrong undermines every pattern and technique built on top.
This page is the entry point for two sections. For content grouped by cross-cutting theme rather than section, use the Concept Map; for design, coordination, and failure-mode patterns, start at Patterns; for standards and surrounding context, start at Reference.
Context Engineering¶
The practice of designing what information enters a model's context window, how it is structured, and what is excluded — to maximize the quality and reliability of agent output. Covers attention mechanics, compression strategies, caching economics, and retrieval patterns.
Instructions¶
How to write, structure, and layer the instruction files that govern agent behavior — from CLAUDE.md conventions and hierarchical scoping to compliance ceilings and negative-space constraints.