AI Assisted Development

Table of Contents

• Workshop
  • Overview
  • What to Expect
  • Getting Started
• Task
• Human-AI Pair-Programming: A Rough Guide
  • IQRE Process
  • Workshop Phases
    • Conception
    • Environment & Tasks
    • Implementation
    • Context Management
    • Presentation
  • Key Guidelines
    • AI Collaboration
    • Quality Assurance
    • Success Criteria

Workshop

Overview

In this workshop, you'll build a travel planning application using Claude Code or Codex as your AI partner. The focus is on practicing effective AI-assisted development, improving prompt hygiene, and fostering human oversight while collaborating with AI.

Getting Started

Create a new repository using this repository as a template.

Task

Details of the task are contained in the brief.

IQRE Process

Follow these four steps consistently throughout the workshop:

1. Iterate: Share ideas/request code from AI and develop specifications or features through iteration.
2. Question: Review AI proposal, identify gaps, and refine through follow-up questions.
3. Accept: If AI proposal is acceptable, allow it to generate the code or specs.
4. Review/Create: Understand generated code/specs. If inspired, create a new, enhanced solution based on AI's output.
5. Explain: Present outputs to teammates, emphasising clear foundations and alignment.

---

Workshop Phases

Note

All prompts referred to in the below section are available here.

CONCEPTION

• Repository Setup: Following Getting Started
• Specification Development:
  • Initialise a new instance of Claude Code or Codex. Use the GENERATE SPECS prompt to have a conversation with the LLM and determine the specifications of your project.
  • At the end of the conversation, use the SPEC WRAP-UP prompt - this should create FUNCTIONAL.md, ARCHITECTURE.md, and CLAUDE.md files.

Output: Initial documentation pushed to repo

ENVIRONMENT & TASKS

Warning

Set up your environment, install your dependencies etc. manually. AI can be terrible at this and using AI for setup could add a lot of config issues to your project before you can even get started.

• Use the GENERATE TO-DO prompt to create TO-DO.md. Remember to follow the IQRE methodology! Check that your tasks actually make sense so that you don't end up with a lot of vague, impossibly scoped tasks that no one could follow!
• Set up environment, frameworks, folder structure, install dependencies
• Review tasks for dependencies and overlaps

Output: Ready-to-code environment with structured to-do list

IMPLEMENTATION

Per Task Process:

1. Use KICKOFF/REFRESH MEMORY prompt
2. Implement features
3. Make sure to review constantly
4. Use CONTEXT RESET after task completion

Between Sessions:

• Update CLAUDE.md with learned standards

Output: Incremental feature completion

CONTEXT MANAGEMENT

• Use HISTORY.md for context summaries
• Reset the LLM's context window after each task
• Maintain clean workspace

Output: Archived context for reference, clean workspace

PRESENTATION

• Demo your project
• Show AI collaboration examples
• Present evolved standards
• Reflect on deliberate architectural decisions

Output: 5-minute presentation with examples and demo

---

Key Guidelines

AI Collaboration

• Explicit Prompting: Always tell the LLM which files to reference (it won't do this automatically)
• Context Management: Use CONTEXT RESET prompt to maintain clarity
• Standards Evolution: Update CLAUDE.md when discovering new patterns

Quality Assurance

• Follow IQRE: Apply the four steps consistently
• Review Obsessively: You need to know everything the AI is generating - read AI Code Review Guide for pointers
• Maintain Standards: Keep CLAUDE.md current and concise

Success Criteria

• Effective AI collaboration patterns
• Evolved standards documented in CLAUDE.md
• Clear architectural decisions

Common Pitfalls

AI code generators often struggle with:

• Anything to do with setting up projects, installing dependencies
• Being too ambitious, agreeing to everything
• Staying inside the scope when working on a task
• Using outdated tech stack, outdated versions of dependencies

Remember: You're the human-in-the-loop. Guide the AI, don't just accept its output.