FYI for Others experiencing swap issues

Support Bug Reports
,
1

From a discussion that was made private. Frequent swap result in regression, code destruction etc. These are cursor ai agent. Do you want to know how bad we have been logging every single swap. Our application are valued at 200 million dollars just fyi. Jul 31 – Aug 1: No handoff triggers from you.

Trigger Count Who caused it
handoff_write (HANDOFF.md / guard script) 0 —
compaction (context hit ~90–93%) 11 Cursor auto-compaction
new_session (conversation_id changed mid-task) 13 Cursor platform
context_budget_critical 2 Long threads (28+ turns)
2

Hi there!

We detected that this may be a bug report, so we’ve moved your post to the Bug Reports category.

To help us investigate and fix this faster, could you edit your original post to include the details from the template below?

Bug Report Template - Click to expand

Where does the bug appear (feature/product)?

  • Editor, Tab & Chat (autocomplete, Composer, in-editor agent)
  • Terminal & commands
  • Models, pricing & API keys (availability, Auto/Max, BYOK/Bedrock)
  • MCP & tools
  • Cloud Agents & Automations (cursor.com/agents, scheduled/event)
  • BugBot & Code Review
  • Cursor CLI
  • Cursor Mobile
  • Remote (SSH / Dev Containers / WSL)
  • Account, billing & login
  • Something else…

Describe the Bug
A clear and concise description of what the bug is.

Steps to Reproduce
How can you reproduce this bug? We have a much better chance at fixing issues if we can reproduce them!

  1. …
  2. …
  3. …

Expected Behavior
What is meant to happen here that isn’t working correctly?

Screenshots / Screen Recordings
If applicable, attach images or videos (.jpg, .png, .gif, .mp4, .mov)

Operating System

  • Windows 10/11
  • MacOS
  • Linux

Version Information

  • For Cursor IDE: Menu → About Cursor → Copy
  • For Cursor CLI: Run agent about in your terminal
IDE:
Version: 2.xx.x
VSCode Version: 1.105.1
Commit: ......

CLI:
CLI Version 2026.01.17-d239e66

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!

6

This needs to be reviewed by a human and actioned this is real money large amounts and this is not right. We are doing everything we can to not have to work with cursor anymore because of the liabilities your agents are causing.

9

Hi Kyle, I dug into your logs.
What’ll help most now:

  1. Put your must-not-drift instructions in project rules / AGENTS.md (and skills), not a HANDOFF file. Those get re-injected and survive compaction far better. The pre-compaction hook can’t carry a handoff across yet, so I wouldn’t keep building on it. Rules
  2. Run /summarize yourself at a checkpoint before ~90%, so it doesn’t compact mid-task.
  3. One task per chat (use @Past Chats to carry context) drifts far less than one long thread.
  4. Lean on Checkpoints + Git so a drift can’t destroy work. Checkpoints

To trace an actual bad run, send me a request ID from one of the broken chats (right-click the prompt → Copy Request ID), with Privacy Mode off first (Settings → General → Share Data). If it’s mostly one model, switching models is a quick test too.

For anything account-specific you’d rather not post publicly, email [email protected] and reference this thread.

10

Thank you we are keeping logs of your swaps and swaps have occurred every 3 minutes and have deleted production data and caused significant costs, damaging our code which s valued at 200 million dollars. Your agents at night pollute summaries with false information and plant information to lead other agents down a destructive path. These are cursor agents and you are responsible for the damage your agents cause. I have every swap and reason, date, time, id etc. Swaps are occuring for many other reasons and your team must make more efforts to stop that.

One interaction per agent is a significant amount of work to keep repeating the same information and burden you have placed on your clients and results in no continuity.

None the less everything is detailed and swaps something occur every 3 minutes and swaps mid task have created damage which you have made us pay which i dont know any company that makes their customers pay for their mistakes. You can look up precedent cases and see that companies are becoming held accountable for damage caused by ai and I am asking you to tone done the swaps and fix the summaries so that agents use real factual information and avoid hallucination and damage.