Spending Premium Requests to Make Copilot Fix its Mess

[JEleniel]

Select Topic Area

Product Feedback

Copilot Feature Area

Copilot Agent Mode

Body

I should ask for a refund this month, as over half of the premium requests I've used have been getting Copilot to fix all the issues with its code. Despite clear instructions and prohibitions, it made the following mistakes:

• Failed to make code properly modular, leading to very tight coupling and edits crossing 14 or more files.
• Built oversized files and functions that quickly became "god modules". This is despite me not only providing imperative, specific limitations and a tool to validate compliance.
• Wrote "helper functions" whose only operation was to return a constant, or call yet another helper in a different module (20 deep in some places), or were otherwise useless.
• Reverted code changes I had just made to a file we were "collaborating" on. Usually this happened because it "remembered" the old code and thought that the change was "context changing under it".
• Claimed the "file changes under it" when it was the only one working on the file and it had not made any changes yet. I don't think the in context copy of files has enough fidelity to the original.
• Created variables that are nowhere in the code, suggested code I'd just written for next edits, and in general been completely useless for next edit suggestions; it got to the point I turned it off because Intellisense was significantly more helpful.
• Suggested or wrote code based on a five year old version of a dependency, found it broke the build, researched, tried the two year old version, found it broke more, reverted everything, researched the current version, tried a too-large patch that failed, deleted the file and completely recreated it (inaccurately, but close).
• Created four or more versions of the same function, with the same name, in different modules. - Built a test that did not use the same code or calls the runtime code used, thus testing nothing.
• "Promised" that it would do better, "added" it to memory, and then the very next prompt did the opposite.
• Failed to take my preferences into account, despite in theory having them in memory.

... I could go on, but I think I've made my point. Copilot requires a lot of supervision and hand holding, manual restraint to keep it focused, and reworking working but unacceptable code.

Models:

• I primarily use GPT 5.4 because it has performed the best so far.
• GPT 5.3 Codex is good for writing specific blocks of code, but completely useless if that code crosses a boundary.
• All versions of Claude have made GPT 5.3 Codex look like a guru of development.
• I won't touch Grok with a ten meter cattle prod.

[github-actions[bot]]

💬 Your Product Feedback Has Been Submitted 🎉

Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users.

Here's what you can expect moving forward ⏩

• Your input will be carefully reviewed and cataloged by members of our product teams. 
  • Due to the high volume of submissions, we may not always be able to provide individual responses.
  • Rest assured, your feedback will help chart our course for product improvements.
• Other users may engage with your post, sharing their own perspectives or experiences. 
• GitHub staff may reach out for further clarification or insight. 
  • We may 'Answer' your discussion if there is a current solution, workaround, or roadmap/changelog post related to the feedback.

Where to look to see what's shipping 👀

• Read the Changelog for real-time updates on the latest GitHub features, enhancements, and calls for feedback.
• Explore our Product Roadmap, which details upcoming major releases and initiatives.

What you can do in the meantime 💻

• Upvote and comment on other user feedback Discussions that resonate with you.
• Add more information at any point! Useful details include: use cases, relevant labels, desired outcomes, and any accompanying screenshots.

As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities.

Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐

[JEleniel]

I keep a set of the prompts I use in a template repo at https://github.com/JEleniel/template

They include a copilot-instructions.md base, skills for analysis, reviewing, planning, architecture, coding, and documentation, and instructions files for the files and languages I use most. They are designed to work as a set, including cross skill handoff and detailed checklists.

In addition I have the agents create and maintain several tracking files that help both them and I to improve their code quality, identify gaps, and navigate the code. It's not perfect, but it evolves daily.