I am experiencing an issue with the latest Cursor update where the AI Agent corrupts Korean text when automatically editing files encoded in EUC-KR.
The Problem: When the Agent performs a task or applies a code change to a file that uses EUC-KR encoding, it seems to process the text as UTF-8 by default. This results in all Korean characters being broken (encoding artifacts/mojibake) once the file is saved by the Agent.
Steps to Reproduce:
Open a project containing files with EUC-KR encoding.
Use the Cursor Agent (Composer or Chat) to make a change to one of these files.
Apply the changes.
The Korean strings in the file become unreadable/corrupted.
Request:
Is there a way to force the Agent to respect the existing file encoding?
Does Cursor have a setting to prioritize files.autoGuessEncoding for the AI Agent’s write operations?
Currently, the only workaround is manually converting every file to UTF-8, which isn’t always feasible for legacy projects. Any guidance or a fix for this would be greatly appreciated.
For AI issues: which model did you use?
Model name (e.g., Sonnet 4, Tab…)
For AI issues: add Request ID with privacy disabled
Request ID: f9a7046a-279b-47e5-ab48-6e8dc12daba1
For Background Agent issues, also post the ID: bc-…
Additional Information
Add any other context about the problem here.
Does this stop you from using Cursor?
Yes - Cursor is unusable
Sometimes - I can sometimes use Cursor
No - Cursor works, but with this issue
The more details you provide, the easier it is for us to reproduce and fix the issue. Thanks!