When are you going to fix Hooks?

Support Bug Reports
,
1

Where does the bug appear (feature/product)?

Cursor IDE

Describe the Bug

See [REGRESSION] Hook response fields user_message / agent_message still ignored in Windows v2.0.77

Steps to Reproduce

See [REGRESSION] Hook response fields user_message / agent_message still ignored in Windows v2.0.77

Operating System

Windows 10/11

Current Cursor Version (Menu → About Cursor → Copy)

Version: 2.3.20 (user setup)
VSCode Version: 1.105.1
Commit: e3fd5c7abddff43abcd46fd93c25e85145200ad0
Date: 2026-01-02T20:39:53.706Z
Electron: 37.7.0
Chromium: 138.0.7204.251
Node.js: 22.20.0
V8: 13.8.258.32-electron.0
OS: Windows_NT x64 10.0.26100

Does this stop you from using Cursor

Sometimes - I can sometimes use Cursor

1 Like
sessionStart hook ignores continue: false
Hooks in 2.4.7 are still not working properly
3

Hey, thanks for following up. I see you already reported this in [REGRESSION] Hook response fields user_message / agent_message still ignored in Windows v2.0.77.

Like you mentioned there on Dec 20, the first part of the bug (hooks not being called at all) was fixed. The current issue is that agent_message and user_message from hooks aren’t being passed into the model context.

The team is working on this, it’s a known issue.

4

I got Claude to look at the CLI - it seems your team somehow shipped a half-finished product by mistake! They implemented only 5 out of the 13 hooks: see Cursor CLI hooks - #5 by Chris_Drake

5

Still not working.

This is wasting hundreds or thousands of user hours as we each all have to independently discover that your documentation promises things that your team broke/never implemented.

Cursor 2.3.41 (Current - 2026-01-21)

| Hook Event | Fires? | agent_message reaches AI? | followup_message works? | user_message shown? | Notes |

|------------|--------|----------------------------|---------------------------|----------------------|-------|

| beforeSubmitPrompt | :white_check_mark: Yes | :cross_mark: NO | N/A | :red_question_mark: Unknown | Fires on user/followup messages |

| beforeShellExecution | :white_check_mark: Yes | :cross_mark: NO | N/A | :red_question_mark: Unknown | |

| beforeMCPExecution | :white_check_mark: Yes | :cross_mark: NO | N/A | :red_question_mark: Unknown | |

| beforeReadFile | :white_check_mark: Yes | :cross_mark: NO | N/A | :red_question_mark: Unknown | Undocumented! |

| afterShellExecution | :white_check_mark: Yes | :cross_mark: NO | N/A | N/A | |

| afterMCPExecution | :white_check_mark: Yes | :cross_mark: NO | N/A | N/A | |

| afterFileEdit | :white_check_mark: Yes | :cross_mark: NO | N/A | N/A | |

| afterAgentResponse | :white_check_mark: Yes | :cross_mark: NO | N/A | N/A | Fires when agent completes |

| afterAgentThought | :hourglass_not_done: Untested | :hourglass_not_done: Untested | N/A | N/A | Need thinking model |

| stop | :white_check_mark: Yes | :cross_mark: NO | :white_check_mark: CONFIRMED 08:35:22 | N/A | ONLY working AI comm channel! |

| beforeTabFileRead | :hourglass_not_done: Untested | :hourglass_not_done: Untested | N/A | N/A | Tab-specific |

| afterTabFileEdit | :hourglass_not_done: Untested | :hourglass_not_done: Untested | N/A | N/A | Tab-specific |

Stop Shipping Unstable Builds
6

This is a critical feature and should be the #1 priority to fix imo. I’m happy to provide logs or anything you need to help move this forward.

7

@deanrie - any word on a fix for hooks?

Everyone else in the world has them!

but your guys took a lunch break about 50% through (5% through for the CLI) and never came back to finish their work!

Is Cursor a public company?

8

This topic was automatically closed 22 days after the last reply. New replies are no longer allowed.