Coding with AI tools

Stage 7 · Lesson 3 of 5

Checkpoints and Git as your safety net

Agents: multi-file changes, permissions and safety

Tools snapshot the files they edit so you can roll back. In Claude Code, press Esc twice with an empty input or run /rewind to restore code, the conversation, or both. In Cursor, restore a from the chat timeline; it reverts files, not messages.

In , restore a checkpoint in the chat. All of them share limits: they track the tool's own file edits, not files changed by terminal commands, databases or outside services, and stopping a request does not undo actions already taken.

So is the real safety net: one , a clean before the agent starts, and a worktree when two tasks run at once.

Common traptreating checkpoints as version control and losing work that only ever lived in a session.

A question to check yourself, an AI practice session and 2 cheat sheet entries come when you add the roadmap to your account. Add it

Resources for this stage

Start with

  • GuideFree
    Building Effective Agents

    Anthropic · Advanced

    Anthropic's clear, free essay on how agents work and when simpler workflows are better

Keep handy

Also worth your time

Mark this stage as not done?