Every developer can push, pull, and commit. But when something goes wrong — a force-push that ate your work, a regression that snuck in months ago, a PR with twelve "fix typo" commits — most developers freeze. The tools that actually rescue you (reflog, bisect, interactive rebase) feel like high-stakes surgery: powerful, but the syntax is gnarly, and the mistakes feel irreversible.

This talk reframes those three commands as one toolkit: Git's recovery layer. We'll walk through three real disasters and the recovery flow for each. Then we'll plug in a custom MCP server that lets an AI assistant drive those same flows in natural language — with previews before anything destructive — so you stop avoiding the commands you most need.

In this session, you will learn to:
- Recognize which of the three tools fits which kind of Git emergency
- Run each one by hand when AI isn't available
- Decide where AI-assisted Git earns its keep — and where it doesn't

From Git Panic to Git Zen, with a little help from AI.