Companion repository for the Code with Antonio video course · Watch on YouTube.
Build an MCP server in Python and use it in Claude Code: tools, resources, prompts and tests, until Claude fixes real bugs from your own issue tracker.
A tiny issue tracker in SQLite, served over MCP; Claude Code reads the issues, fixes the bugs in a real project, runs the tests and closes them.
| # | Lesson | Video | Folder |
|---|---|---|---|
| 1 | What is MCP? | 0:00 | lessons/01-what-is-mcp |
| 2 | Your first MCP server | 2:42 | lessons/02-your-first-mcp-server |
| 3 | Tools that do real work | 6:08 | lessons/03-tools-that-do-real-work |
| 4 | Plug it into Claude Code | 8:54 | lessons/04-plug-it-into-claude-code |
| 5 | Resources and prompts | 11:35 | lessons/05-resources-and-prompts |
| 6 | Test it and debug it | 14:29 | lessons/06-test-it-and-debug-it |
| 7 | Share it with your team | 16:53 | lessons/07-share-it-with-your-team |
- MCP Python SDK 2.3.0 (
MCPServer) · spec MCP 2026-07-28 · MCP Inspector 2.9.0 · Claude Code 2.1.288 · uv 0.12.22 · Python 3.12 · pytest 9.1 · Ubuntu (Linux) - Python 3.12+ and uv
cd app
uv sync
uv run pytest
uv run mcp dev server.py # the MCP Inspector (needs Node.js): create a few issues with create_issue
cd demo/textkit
# edit .mcp.json: put the absolute path to app/server.py, then:
claude # approve the tracker server, then ask: What's open in the tracker?README.md this file
lessons/ one folder per lesson: goal, timestamp, commands and files exactly as in the video
app/ the final working project
assets/ thumbnail
MIT — use it for anything, credit appreciated.
Generated on 2026-10-03 from the course scripts. Code with Antonio · AI tutor.