32dots HEIDELBERG AI
Curriculum › GitHub Copilot course › copilot-03
Session 3 medium

Agent Mode: multi-file tasks

LESSONLesson 3 · ~25 min

🎯Goal. Hand Copilot a longer, multi-step task and let it plan, write, and iterate across several files.

▶ Try this prompt

Split my analysis script into a reusable data-loading module and a plotting module, add type hints throughout, and write a quick test for the loader.

Switch the chat picker to Agent mode, then send the task. Copilot proposes changes across multiple files for you to review.

  1. 1In the Copilot Chat panel, change the mode dropdown to Agent. Now Copilot can plan, write, run, and iterate across multiple project files instead of editing one block.
  2. 2Send a task that spans several files. Copilot lists the files it changed, applies edits, and can run code to check its own work.
  3. 3Review before you keep it: read each changed file, then click Keep to accept or Undo to discard. Agent Mode can over-reach, so verify every file it touched.
GitHub Copilot Agent Mode showing a '4 files changed' summary with Keep and Undo buttons and an Agent / Auto mode picker
Agent Mode edits across multiple files at once — here it changed 4 files and offers Keep or Undo so you review every edit before accepting. Source: https://github.com/features/copilot

✓You'll see. Copilot reorganising your code across several files at once, with a 'files changed' summary and Keep / Undo controls to review its work.

💳Cost. Agent Mode consumes AI Credits. Free includes limited agent usage; Pro ($10/mo) comes with $15/mo in AI Credits, Pro+ ($39/mo) adds premium models plus $70/mo in Credits. A deep multi-file run costs more than a single chat.

💡Takeaway. Agent Mode tackles work too big for autocomplete — but you stay the reviewer, keeping or undoing each change.