Hands-on Coding Assistants

A Developer's Mental Model for AI-Assisted Engineering

AI coding tools can write code for you. But writing code was never the bottleneck — making the right decisions, catching bad suggestions before they compound, and maintaining quality across a real codebase is. Most developers use these tools on autopilot. This series shows you how to use them with intention.

Across 17 chapters and 6 modules, you'll learn a repeatable mental model — Plan, Delegate, Review, Correct (PDRC) — and apply it hands-on to every stage of the development cycle: from writing your first prompt to shipping a full feature with multi-agent collaboration, automated code review, security hooks, and measurable impact.

What you'll walk away with

• A mental model for working with AI that prevents the "accept all" trap and keeps you in control of quality.
• Custom agents, instruction files, and skills tailored to your codebase — so the AI knows your conventions before you even type a prompt.
• Practical workflows for test generation, code review, debugging, refactoring, and documentation — the daily tasks where AI delivers the most measurable value.
• MCP integrations and hooks that extend what the AI can access and automate, from external APIs to pre-execution security validation.
• Governance, security, and impact measurement strategies so you can adopt AI tools at team and org level with confidence.
• An end-to-end final project where you run the complete PDRC cycle on a real codebase.

Every chapter is hands-on: real repos, real tools, real trade-offs. No filler, no hype — just the engineering discipline that makes AI assistants actually useful.

---

Module 00 — Foundations & Philosophy

Core concepts, the PDRC mental model, setup, and prompt engineering.

• Ch 0: Before You Start
  Prerequisites, tools, accounts, and knowledge you need to follow this series effectively.
• Ch 1: Introduction & Overview
  Code assistants vs. code agents, tool landscape, where AI delivers proven value, and what AI does NOT do well.
• Ch 2: The Mental Model: Plan, Delegate, Review, Correct (PDRC)
  The way I think about working with coding agents — a personal adaptation of the PDCA cycle for AI-assisted development.
• Ch 3: Setup & Practical Integration
  Configuring Copilot/agents in the IDE, repository connection, custom instructions, MCP servers, and the new standalone Copilot CLI.
• Ch 4: Prompt Engineering for Code Assistants
  From basics to advanced prompt structures, context management, and strategies for complex tasks.

Module 01 — Agent Customization

Custom agents, AGENTS.md, repository instructions, and agent skills.

• Ch 5: Custom Agents & Sub-Agents
  Building specialized agents with .agent.md for testing, planning, documentation, and security review.
• Ch 6: AGENTS.md & Project Context
  Using AGENTS.md as a standard for guiding agents with build steps, conventions, and architecture.
• Ch 7: Repository Custom Instructions
  Repository-wide and path-specific instructions with copilot-instructions.md and .instructions.md files.
• Ch 8: Agent Skills & Modular Expertise
  Building and using Agent Skills with SKILL.md and scripts for conditional, modular agent capabilities.

Module 02 — AI in the Daily Development Cycle

Test generation, code review, debugging, refactoring, and documentation.

• Ch 9: Test Generation & Improvement with AI
  AI-assisted TDD, generating unit/integration/E2E tests, and coverage gap analysis.
• Ch 10: AI-Assisted Code Review
  Copilot Code Review, custom review instructions, automatic reviews on PRs, and implementation suggestions.
• Ch 11: Debugging, Refactoring & Iteration with AI
  AI-assisted debugging, legacy code refactoring, codebase onboarding, and iteration strategies.
• Ch 12: Automated Documentation
  Code documentation generation, API/schema docs, AI-assisted commit messages, and PR descriptions.

Module 03 — Extensibility & Integrations

MCP (Model Context Protocol), hooks, validation, and agent automation.

• Ch 13: MCP (Model Context Protocol)
  Open protocol for connecting AI to external systems — configuration, practical examples, and security.
• Ch 14: Hooks, Validation & Agent Automation
  Execute shell commands at key agent execution points, setup steps, and integration with project tools.

Module 04 — Security, Governance & Impact Measurement

AI-first security, governance, IP policy, impact metrics, and continuous improvement.

• Ch 15: AI-First Security & Governance
  Risks, agent firewalls, separating agent access, testing/releasing custom agents, and IP considerations.
• Ch 16: Impact Measurement & Continuous Improvement
  Defining metrics, developer experience data, analyzing agent performance, and spec-driven development.

Module 05 — Full Cycle

End-to-end final project applying the complete PDRC cycle on a real codebase.

• Ch 17: End-to-End Final Project
  Complete challenge applying the full PDRC cycle with multi-agent collaboration on a real codebase.
• Ch 19: References
  All external references cited throughout the Hands-on Coding Assistants series, organized by category.

---

License & Attribution
This series content is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0) . You are free to share and adapt this material for any purpose, including commercial use, as long as you give appropriate credit.

Please cite as: "Hands-on Coding Assistants" by William Oliveira — woliveiras.com