Course Outline
09:00 – 09:15 — Welcome & Overview
- Overview of the day’s objectives, technical setup verification, and introductions
09:15 – 10:00 — Module 1: AI Code Quality & the Plan–Execute–Review Principle
- Five key factors influencing outcomes: model selection, codebase context, tooling, prompt precision, and workflow design
- The advantages of separating planning, execution, and review phases over relying on a single prompt
- A decision-making framework to guide the remainder of the workshop
10:00 – 10:45 — Module 2: From Autocomplete to Agent: The Claude Code Paradigm
- Distinguishing Claude Code from tools like Copilot and Cursor, and understanding why this distinction is critical
- The agentic cycle: read, plan, act, verify
- Live demonstration: executing a multi-file task via a single instruction
10:45 – 12:15 — Module 3: Setup & Personalisation: CLAUDE.md, Skills, Hooks, Agents & MCP
- CLAUDE.md: encoding project memory, architectural guidelines, and team conventions
- Skills and hooks: creating reusable slash-command workflows and event-driven automations
- Sub-agents: understanding how Claude Code internally delegates parallel tasks
- MCP: connecting external tools — such as GitHub, databases, and internal APIs — while maintaining security boundaries
- Hands-on session: configuring Claude Code and customising your personal setup
12:15 – 12:45 — Lunch Break
12:45 – 13:45 — Case 1: Writing Code with Claude Code
- Delegating implementation tasks with clear context and acceptance criteria
- Iterating on results: knowing when to accept outputs and when to redirect them
- Hands-on session: building a real feature from start to finish
13:45 – 14:45 — Case 2: Doing Code Review with Claude Code
- Structuring review delegations focused on security, design, and performance
- Conducting multi-file reviews with traceability — tracking changes, understanding rationale, and determining verification steps
- Hands-on session: reviewing actual code from participants
14:45 – 15:45 — Case 3: Designing with Claude Code
- Generating architectural decisions and Architecture Decision Records (ADRs) based on requirements
- Exploring design alternatives and identifying potential trade-offs
- Hands-on session: designing a component or feature using Claude Code
15:45 – 16:15 — Recap & Q&A
- Summary of key takeaways and commitment to three actionable changes for each participant
- Open Q&A session
Requirements
- Proficient experience with at least one programming language
- Comfortability working within a terminal or command-line environment
- Fundamental Git knowledge, including commits, branches, and pull requests
- A laptop with reliable internet access — a Claude Pro account is mandatory (registration instructions will be provided)
Testimonials (3)
Using Claude Code in a more efficient way
Virgil Trif - Frequentis
Course - Claude Code: Agentic AI Development · 1-Day
claude mcp, how to promote and sell AI created software
Precup Andrei - Frequentis
Course - Claude Code: Agentic AI Development · 1-Day
The workshop to build an app
