Copilot doesn't complete, 'Applying edits' It seems to get stuck and it often will not use the added context

[Pneumanifest]

Select Topic Area

Bug

Copilot Feature Area

VS Code

Body

I am a member of an organization account, and over the last couple weeks there has been a persistent problem where the chat will not complete the edits in VScode. It shows applying edits, some 99% others 100% and some just sit at generating edits. It can sit like that for hours if I let it. If I hit stop it will show many of the edits it made but code is heavily truncated.
It also keeps telling me to provide the files for reference even though I already used the add context or have the file open in the editor. Then proceeds to say that it is unable to read files that I add. As in its not allowed to do that kind of task, and it tells me to copy paste the code in the chat instead.
What has changed? It was so useful and did edits/fixes so easily on the same project just a week prior, without any issues.

[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. ⭐

[metinkoyuncuu]

I have a same problem

[umbra99]

Having the similar issue editing files just gets stuck every 2-3 edits. It happens independent of the model used and It does use premium requests if the model is a premium model.

[StackingLayers]

It does use premium requests if the model is a premium model.

This is also a concern. Lost of wasted premium points/tokens due to tasks not completing.

[foxalabs]

Same issue, it's ok at initially generating the files, but when it edits the files it will get to 100% and just hang and never return Using Agent mode and Claude 4

[Ash645]

same here as well, it's frustrating at times as closing VS Code is the only fix, and I lose edits and it has to re-read everything that happened right before to get to this point..

[mechanicmuthu]

Same with me. Stuck at different percentages. Nothing is saved. Same is counted.

[Kaboniro]

so far what I'm noticing is for large projects the model might be somehow confusing contexts or have a large cache of contexts impacting its behavior or something i think, so my workaround is go to the specific file i need a fix and address the issue there one by one let it generate the fix and i apply it manually

[mrarejimmyz]

I think the same , and when its fixing the issues we need to also see what is it trying to do and stop before it creates a completely new problem.

[dhit-mdavis]

yep, specifying the exact files, even in the very small, starter projects I have, has been helpful

A super helpful trick has been to make a text file of what I want or what my target is and then always include that file for context as I attack a specific problem or new feature

[mrarejimmyz]

It's a caching issue—whenever it happens, I simply restart my PC and it resolves the problem.

[cyril-accenta]

Same here, from time to time it gets stuck as described above: “Applying edits (100%)” (the percentage may vary).

[itzi-vignesh]

Same issues, but after 80 to 90% of applying the edits, the error occurs, refreshes the whole application, and directs to retry

I have also raised a token regarding this issue, https://github.com/orgs/community/discussions/165430

[gustavoaca1997]

[foxalabs]

I've noticed it CAN affect files of any size, but it gets worse on large files.. that may just be observer bias, but that seems to be the vibe I get. Also worse when the models are busy, i.e. US work hours.

[aniketan]

I am facing the same problem; I hope the GitHub team will resolve it soon.
One trick I use when it gets stuck,
Pause it and type "provide a complete updated file, and I will copy and paste it."

[andrewhaller]

This is complete horse shit. GiHub refund us until you can provide a WORKING, NON-FREEZING service. Somebody needs to be kicked really hard for this.

[shmed]

Same thing is happening for me pretty often. I usually press the "stop" copilot button and ask the AI to try again, but its frustrating to lose credits, and more importantly, it breaks very long "agent" operations and then when you restart it loses a lot of context and it can have a hard time pick up from where it was.

[thegrince]

Same for me. In Edit mode, "Generating edits..." hangs.
Sometimes at 99%, sometimes at the beginning, medium project, very small file, simple edit like 'add database table check before saving data'

[tgv1975]

Feel free to downvote, but I need this off my chest: Never in my entire career have I seen such an appalling integration, full of errors, often downright dangerous, for a paid service.

I get that it's early and the product needs to get to market on the current AI momentum, but man, Copilot in VS Code absolutely reeks.

So yes, I too run into the issues the OP mentions.

[dhit-mdavis]

I'm not affiliated with Microsoft or Github by any means BUT I also don't use CoPilot for VS Code, I use Visual Studio. I've been having great results but not perfect by any means.

I don't care about Claude vs OpenAI

Agreed that the human brain and spirit should not be worried at this time but it doesn't need to be perfect to be great because with enough time, it'll get closer and closer

[maxmnetwork]

completely agreed about it doesn't need to be perfect to be great, but i must say that I have corrected Copilot, J!dev, JoomGPT and other models' mistakes more often than I should, and given that i'm not skilled or trained in coding, the fact that I'm told "You're absolutely right" by AI more than any human has ever told me is pretty darn scary. If i wasnt as sharp as I am and allowed Copilot or other AI models make decisions for me, I would never get my project completed because of all the wrong code it gives me and disobedience to do it correctly despite it being a purported 'expert' coding model. I'm looking forward to advancements, but they need to be quite significant advancements to impress me at this point.

[dhit-mdavis]

Oh yeah, completely agreed on that Max. It's very scary dealing with this technology. It's exciting as hell but the more you think about it being at the root of so many important decisions or sources of information, it definitely makes you cautious. I still make sure I manually click allow on all requests, I still review the code before saving it. Something feels very high potential about it though and I can't escape that feeling like I could do way more in a period of time (not a short period always but sometimes, it can be impressively fast, at least for a brand new project vs a an existing, large scale project)

[maxmnetwork]

I have also used AILawyer ChatGPT to help me write some legal papers (I'm a trained paralegal in a pro se case against my former lawyer-supervisor) which has been incredibly helpful for brainstorming and research, but I've had to discipline and correct the model for completely fabricating facts and caselaw. It completely fabricates nonexistent law and facts with an authoritative voice and posture. If I was a common person thinking the models are always correct, and used its legal and factual references without verifying them myself, i would be laughed out of court with sanctions for wasting the court's time. I just read about an attorney who did exactly that - she used AI to write her legal arguments and cite caselaw that was completely fabricated, and she took no effort to verify any of it, so when the judge searched for the caselaw and couldn't find it, need I say the judge was P.O'd and the attorney was in deep deep trouble... Pretty scary when your attorney is relying on AI without verification ...

[dhit-mdavis]

That's exactly where my concerns are. The output has already infiltrated many ways of life and many industries and if you're visually conscious and aware, you have to start thinking about how any Apps you make or interact with are affected by that unpredictability and unreliability.

[simone-cominato]

Suddenly it happened to me too, but in my case signing out from the Github account (Accounts icon in the lower left corner in VS Code) and then signing back in solved the issue.

[CHMEX]

same issue here and it doesnt edit anymore

no solution so far is that maybe an outage?

[simone-cominato]

Did you try to sign out and back in as suggested in my previous comment?

[CHMEX]

yes that didnt work but what did work was deinstall the complete VSCode and reinstall it it also made super faster so that's also a win now all working again

[PRIYAtechky]

It sounds like you’re running into a few issues with the VS Code integration:

1. Edits Stuck at 99% or 100%

   • This is often related to network lag or a temporary backend issue. Try restarting VS Code and ensuring you’re on the latest version of the GitHub Copilot or Chat extension.

2. Code Truncation After Stopping

   • When you stop generation mid-way, partial edits are shown. To avoid truncation, let the edit complete fully, or split large edits into smaller chunks.

3. "Provide Files for Reference" Issue

   • Even if files are open, sometimes the extension loses context.
   • Workaround: Use “Add Context” again or manually copy-paste the code.
   • Make sure the workspace trust is enabled so the extension can read project files.

4. "Not Allowed to Read Files" Error

   • This usually happens if the project folder isn’t fully trusted in VS Code.
   • Go to the top bar → File > Manage > Workspace Trust and enable trust for the repo.

5. What Changed Recently

   • GitHub is actively updating the Copilot/Chat extensions. Some stricter rules have been introduced around file access for security reasons.
   • This means tasks that previously worked automatically might now require explicit context or user confirmation.

Next Steps

• Update VS Code and all related extensions.
• Check workspace trust settings.
• Break down edits into smaller pieces.

[DevNullInc]

I'm personally getting tired of it parsing the entire damn file (sometimes over 50,000 lines of code to update one function.

[Doommarine1991]

well copilot keeps giving me half - assed responses! what i mean is ive told it to take the full code 1700 lines and merge it / replace a part of it and then spit out the fixed one but it never ever does! its given me small chunks of parts that are random and its full on stupid

[MikeRMTS]

Mine just started to do the same this morning. Very weird. All file changes come back as just partials.

[hmexdev]

getting the same today

[Baragon2021]

Edit mode is not reliable. 10-15% of times will recommend what seem to be incomplete or truncated code modifications , especially within larger files or for more extensive modifications (where the edit mode would actually be most useful). Edit mode does allow you to scroll through proposed edits but it does not allow you to selectively apply individual edits, you must pick all or none. Also, it slowly parses a 5000 line file just to add a new 20 line function. i know that manual copy and past would be quicker, but then what is the point of edit or agent mode. I am getting 95% good advice from claude sonnet 4.5 but then copilot edit mode is only applying these recommendations correctly 85-90% of time = overall accuracy ~80-85%. Unacceptable for a paid service.

[Baragon2021]

Compared to 3 months ago, edit mode is getting more accurate and somewhat faster to parse a large file but still only gives correct applications of edits recommended by model 85-90 percent of time for large files (>1000 lines) so be careful to check it against the recommendation. I often use ask first (to examine the proposed modifications) and then switch to edit and ask edit to make the recommended modifications made in the ask chat one at a time. This seems to work better at times.

[Baragon2021]

i think i will go back to using ask mode and try "apply in editor" and see if that works more reliably

[tjohnson4]

I get into this state way to often and it is super frustrating. The chat will tell me everything it will do and say it will make the changes and then never does. Then I need to tell it to proceed but it needs to reevaluate my last command and then reasses the plan and then tellls me again what it will do. Such a waste of tokens and my time and polutes the context/session.

Github Co-pilot
Mode: Agent
Model: Gemini 2.5 Pro
VSCode
Version: 1.105.1 (Universal)
Commit: 7d842fb85a0275a4a8e4d7e040d2625abbf7f084
Date: 2025-10-14T22:33:36.618Z
Electron: 37.6.0
ElectronBuildId: 12502201
Chromium: 138.0.7204.251
Node.js: 22.19.0
V8: 13.8.258.32-electron.0
OS: Darwin arm64 24.6.0**

[TheSumitBanik]

Same here.

[BenoMosko]

any solution?
still experiencing this issue.

[Neil-Crago]

Same problem

[oregonpillow]

The fix for me on a Mac was a re-install of VS Code.

1. Recommended: ensure backup is enabled. (backup your current settings / extensions to cloud)

2. Uninstall VS-Code

3. Re-install
   brew install --cask visual-studio-code

4. copilot editing working again