Chapter 01
Context engineering fundamentals
Keep an agent focused by balancing relevance, structure, and context cost.
AI-ASSISTED DEVELOPMENT · DEVELOPER HANDBOOK
Working with AI Agents
Developer Handbook
Nine field guides on context engineering, agent capabilities, specification management, and the day-to-day craft of shipping code with an agent.
- Author
- Mykhailo Markhain
- Edition
- Version 1
- Format
- 33-page PDF
Inside the handbook
Nine field guides for agentic development.
Read the chapters in order, or jump straight to the capability, workflow, or team practice in front of you.
Chapter 02
Extending agent capabilities
Understand context files, tools, prompts, skills, subagents, and hooks.
Chapter 03
Choosing the right capability type
A practical decision process for selecting the right extension surface.
Chapter 04
Model Context Protocol guidance
Evaluate MCP servers by status, configuration, and context cost.
Chapter 05
Spec-driven development
Write focused requirements and design documents that agents can use.
Chapter 06
Specification management
Separate system truth from task intent and keep both useful over time.
Chapter 07
From specification to implementation
Use the explore, plan, code, and commit loop to ship with discipline.
Chapter 08
Making code changes with an agent
Build feedback loops, checkpoints, and review habits into everyday work.
Chapter 09
Team learning approaches
Turn experiments, champions, and retrospectives into shared capability.
Who it is for
Builders shipping with AI agents.
This developer handbook is for engineers, technical leads, and teams using coding agents to explore repositories, write specifications, change code, and learn from the work. It keeps the focus on practical context, clear boundaries, and feedback loops that make agent-assisted software safer to ship.
AI agentsagentic developmentcontext engineeringMCPModel Context Protocolspec-driven developmentAI-assisted development