Is there a way to get Claude to reliably automatically fix IDE problems in

Support Help
1

Has anyone figured out a way to get Cursor/Claude to reliably fix lint errors / IDE problems as it goes?

I used basedpyright, which surfaces problems in the Cursor IDE “Problems” UI, and it seems like

  1. The only way Claude 4.0/4.5 can interact with basedpyright is the command line (rather than MCP, which is what I’d expect from the “Auto-Fix Lints” feature?)

  2. I have to remind it constantly to fix the IDE problems

2

In every project, at the root, I ask Agent to write a PowerShell script that will run linters and output their results. I tried to synthesize my approach in Agent Enforcer, but so far it’s broken and nobody needs it.

3

Thanks @Artemonim !

It does seem like Cursor should make your approach not be required :slight_smile:

Linting turns coding further into a verifiable domain, so it’s nuts that Cursor can’t reliably use linter errors as part of its completion verification loop.

4

For example, in an Android project, I have a local CI/CD PowerShell script with 1300 lines of code, which Gemini classifies as a senior-level script :grin:

There is already a linter inside Cursor Agent, but it is weak.

5

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.