Edit a large notebook with Copilot Agent Mode
The author asked Copilot Agent Mode for moderate changes and basic features in a large ipynb; it got halfway, stopped mid-line, and broke the notebook. ·
Done withGitHub Copilot
- What they did
- A commenter described using Copilot Agent Mode in PyCharm on a fairly large Jupyter notebook, about 3,000 lines, to make moderate edits and add basic features. The agent would start rewriting the notebook, get partway through, then stop mid-line and corrupt more than half the file; the person reverted the changes and tried several models before giving up.
- How it went
- It failed repeatedly on the same notebook, breaking the file badly enough that the author reverted each time and stopped using Agent Mode for that job. They switched to Gemini in chat mode instead, saying it worked better for AI coding.
- Worth knowing
- The main catch was file safety: on this large ipynb, the agent repeatedly stopped halfway through edits and damaged over half the notebook.
Try it yourself with GitHub Copilot
I have a large notebook, [file name], and want these changes: [moderate change 1], [moderate change 2], and a basic feature to [feature]. Work one cell or section at a time, validate that the notebook is still valid JSON and runs after each edit, and never stop partway through a cell. When done, list what changed per section, and ask me before making any big restructuring edits.
Want help getting this working? Join the Vibe Coding Academy ↗ Run by Alex Finn, who also runs this site.
1 more account of this
Source: Reddit · Jun 3, 2025
Related use cases
Run a solo company with eight agent departments
A founder set up CEO, CFO, COO, lawyer, accountant, marketing, CTO, and improver agents that share memory and consult each other.
Build a full-stack project from scratch in Copilot agent mode
The poster built a full-stack project with Copilot agent mode using Claude 3.7 Sonnet; it needed babysitting through setup problems but eventually produced styled, working React components.
Break development work into phases for Copilot
The poster splits work into phases with explicit tasks so Copilot's agent works through development steps instead of everything at once.
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Build a full CRM with email, video and calendar
A community member used Replit Agent over a couple of weeks to build a CRM with email integration, video sales letter recording and meeting scheduling.
Build an RFP analysis tool
The author built a simple RFP analysis tool with Replit Agent, using a Flask backend and an HTML, Tailwind and JavaScript frontend.
Get 5 like this every morning
The best things people got an AI agent to do, each with the prompt to try it.