LC

lucab85/claude-code-bootcamp

Developer tools
12 stars Quality 56 Trend 12NOASSERTION

Claude Code Extended: Build 10 Real-World Projects with Claude Code

Overview

· part of A 5-hour live virtual bootcamp by — Claude Code Certified Expert Instructor · AI-first Developer · Author & Educator 30 May 2026, 09:00 AM EST · Live, Virtual & Practical → Open exercises/part-01/README.md. Everything else on this page is reference material. (complete before the live session — ~30 min): - [ ] installed and signed in (any tier) - [ ] on PATH (primary track) - [ ] on PATH (secondary track for parts 2/4/5) - [ ] on PATH - [ ] An IDE you can drive (VS Code recommended) - [ ] macOS / Linux / Windows-via-WSL2 — native PowerShell is supported , work through parts 01 → 10 in order. Each part is a ~22-minute block with a slide deck + an exercise + a reference solution. Part 11 is the closing Q&A + exam briefing. Most developers use AI as autocomplete. This bootcamp teaches you to use as a development partner — one that plans, codes, tests, reviews, refactors, and documents alongside you.

README

Claude Code Bootcamp — Build 10 Real-World Projects with Claude Code

Packt Certification · part of LLM Engineering by Packt A 5-hour live virtual bootcamp by Luca Berton — Claude Code Certified Expert Instructor · AI-first Developer · Author & Educator Inaugural delivery: 30 May 2026, 09:00 AM EST · Live, Virtual & Practical


Start here

Where do I start? → Open exercises/part-01/README.md. Everything else on this page is reference material.

Pre-work checklist (complete before the live session — ~30 min):

  • [ ] Claude Code installed and signed in (any tier)
  • [ ] Python 3.11+ on PATH (primary track)
  • [ ] Node.js 20+ on PATH (secondary track for parts 2/4/5)
  • [ ] Git ≥ 2.30 on PATH
  • [ ] An IDE you can drive (VS Code recommended)
  • [ ] macOS / Linux / Windows-via-WSL2 — native PowerShell is not supported

Full pre-work is in student-guide.md.

During the bootcamp, work through parts 01 → 10 in order. Each part is a ~22-minute block with a slide deck + an exercise + a reference solution. Part 11 is the closing Q&A + exam briefing.

flowchart LR
    A[Pre-work~30 min] --> B[Bootcampparts 01–104 h instruction]
    B --> C[Part 11Q&A + exam briefing~30 min]
    C --> D[Packt Certificate+ Skills library]

Table of Contents


Overview

Most developers use AI as autocomplete. This bootcamp teaches you to use Claude Code as a development partner — one that plans, codes, tests, reviews, refactors, and documents alongside you. Across 10 hands-on projects in a single afternoon you will ship a CLI, a REST API, a test suite, a UI from a wireframe, a refactor with handoff docs, a Git workflow, a personal skills library, and a production readiness review.

You leave with: a portfolio of 10 deliverables, a reusable Claude Skills library (10+ skills, MIT-licensed), and a Packt Certification Certificate of Completion.

Audience

  • Beginner-to-intermediate developers
  • Tech leads adopting AI-assisted workflows
  • DevOps and automation engineers
  • AI-first builders moving from prompts to shipped software

Prerequisites

  • Basic programming literacy in any language
  • Git basics (clone, branch, commit, push)
  • Working Claude Code access (any tier)
  • macOS / Linux / Windows-via-WSL2 — native PowerShell is not supported
  • Python 3.11+ (primary track) and Node.js 20+ (secondary track for modules 2/4/5)
  • ~30 minutes of mandatory pre-work completed before the live session — see student-guide.md

Learning outcomes

By the end of the bootcamp you can:

  1. Drive Claude Code through plan → implement → test → review → commit loops
  2. Write a CLAUDE.md brain file that gives Claude project context
  3. Generate, evaluate, and select Best-of-N implementations
  4. Generate test suites and detect AI-generated bugs
  5. Run AI-assisted Git feature-branch workflows safely
  6. Convert a wireframe into a working UI via multimodal prompts
  7. Refactor a module under constraints and ship handoff docs
  8. Author reusable Claude Skills and slash commands
  9. Produce a 5-axis production readiness report
  10. Self-review AI output against a Code Review Rubric you author

Schedule

5 hours total = 4h instruction (240 min) + 1h breaks/Q&A/exam briefing (60 min).

# Module Time (min) Deliverable
1 Welcome, Setup & AI-First Mindset 20 AI Coding Workspace
2 Prompting Like a Tech Lead 24 CLI Task Manager
3 Project Context with CLAUDE.md 22 Project Brain File
4 Build Faster with Best-of-N 30 Notes App API
5 Testing, Debugging & Self-Review 28 Tests + Bug Fixes + Code Review Rubric
6 Git Workflows for Safe AI Dev 22 Feature Branch Workflow
7 Multimodal: Screenshot to UI 30 Dashboard UI
8 Refactoring & Documentation at Scale 24 Refactor + Handoff Docs
9 Skills, Hooks, MCP & Plugins 22 Personal Claude Skills / Command Library
10 Production Readiness 18 Production Readiness Report
Instruction total 240
11 Q&A, exam briefing, next steps (closing block) 30 Action plan for Monday
Breaks + buffer 30 —
Schedule total 300

Projects

The 10 projects students build, in order:

  1. AI Coding Workspace — repo + CLAUDE.md + first guided Claude Code run
  2. CLI Task Manager — Python primary; Node.js secondary
  3. Project Brain File — a real CLAUDE.md for an existing repo
  4. Notes App API — Python (FastAPI + SQLite); Node.js (Hono + better-sqlite3)
  5. Tests + Bug Fixes + Code Review Rubric — pytest + 2 bug fixes + a student-authored rubric
  6. Feature Branch Workflow — feature branch + AI-generated commit + PR text
  7. Dashboard UI — single-page UI rendered from a wireframe (Python)
  8. Refactor + Handoff Docs — refactor a module + HANDOFF.md + ARCHITECTURE.md (Python)
  9. Personal Claude Skills / Command Library — author at least one new SKILL.md, and learn how skills, hooks, and MCP servers package into an installable plugin
  10. Production Readiness Report — 5-axis report on a prior project

Module 5’s “Code Review Rubric” is the student-built deliverable at exercises/part-05/code-review-rubric.md. The instructor grading rubric lives at assessments/rubric.md. They are different artifacts.

Repository layout

.
├── README.md                       # this file
├── LICENSE                         # CC BY-NC-SA 4.0 (course materials)
├── instructor-guide.md             # live-delivery + grading guide
├── student-guide.md                # student onboarding + pre-work + submission
├── certificate-template.md         # Packt Certification template
├── slides/                         # 10 Marp decks + build script
│   ├── README.md
│   ├── deploy-pptx.sh
│   └── part-01-…-part-10-….md (+ part-11 closing block)
├── exercises/                      # 10 hands-on labs (parts 01–10)
│   └── part-01/ … part-10/
├── skills/                         # MIT-licensed Claude Skills library (12 skills)
│   ├── LICENSE
│   ├── README.md
│   └── /SKILL.md
├── assessments/                    # quiz, practical, reflection, rubric, key
├── scripts/
│   ├── preflight.sh                # 15-gate pre-cohort audit
│   ├── check-slide-overflow.sh
│   ├── check-contrast.sh
│   └── check-verbatim-blocks.sh
├── archive/                        # off-agenda content (optional warm-up)
└── specs/                          # Spec Kit feature specs (maintainer-facing)

Build

Build all 11 slide decks (10 modules + closing block) to PPTX (and optionally PDF/HTML):

cd slides
./deploy-pptx.sh            # PPTX only → slides/dist/pptx/
./deploy-pptx.sh --all      # PPTX + PDF + HTML → slides/dist/{pptx,pdf,html}/
./deploy-pptx.sh --pdf      # PPTX + PDF
./deploy-pptx.sh --clean    # remove dist/ before rebuilding

All decks build to a single flat tree under slides/dist// using the self-contained wow-beginner theme. See slides/themes/README.md.

The script auto-detects a global marp and falls back to npx --yes @marp-team/marp-cli@latest. Set CHROME_PATH if Marp can’t find Chromium for PPTX/PDF export. See slides/README.md.

Run the audit

Before every cohort, run the pre-flight audit to catch broken cross-links, missing sections, drifted slide artefacts, and stray [NEEDS CLARIFICATION] tokens:

bash scripts/preflight.sh           # all 15 gates; RC=0 means safe to deliver
bash scripts/preflight.sh --quick   # skip the slow slide-overflow render
bash scripts/preflight.sh --gate audit.cross-links --verbose

Gate reference: instructor-guide.md § Pre-delivery audit.

Assessment

Component Weight Artifact
Knowledge quiz 40% assessments/knowledge-quiz.md
Practical task 40% assessments/practical-task.md
Code review reflection 20% assessments/code-review-reflection.md
Pass threshold 70% assessments/rubric.md

Students who achieve ≥70% receive a Packt Certification Certificate of Completion issued from certificate-template.md.

Submission workflow: students zip their deliverables (module-00-prework/ + module-01/ … module-10/ + assessments/) and upload to the Packt LMS. Instructor grades locally — see instructor-guide.md.

License

This repository uses a dual-license scheme:

  • Course materials (this README, slides, exercises, guides, assessments, certificate template, scripts) — CC BY-NC-SA 4.0. You may share and adapt for non-commercial use with attribution.
  • skills/ directory — MIT. Graduates may reuse skills in commercial projects without further restriction.

“Packt Certification” and “LLM Engineering by Packt” are trademarks of Packt Publishing Ltd., used here under endorsement.

Optional pre-bootcamp warm-up (archived)

There is an earlier beginner track kept under archive/beginner/. It is not part of the 4-hour bootcamp and is not maintained on the May 2026 refresh path. Use it only as a self-study warm-up if you have never used Claude Code before. Instructors: do not point students here during a cohort.


Instructor: Luca Berton — Claude Code Certified Expert Instructor · AI-first Developer · Author & Educator Endorsement: Packt Certification · LLM Engineering by Packt

View the full README on GitHub

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add lucab85/claude-code-bootcamp
New to MCP? Start with the mcpskill.org guide →