Regression in Adobe Acrobat Chrome extension - verified and reproducible

We believe we've identified a regression in the Adobe Acrobat Chrome extension introduced in the recent 26.7.x releases. 

Our web application generates PDF reports in response to an authenticated HTTP POST. 

The server returns the PDF successfully. However, when the Adobe extension is enabled, immediately after the successful POST, the browser performs an unexpected GET request to the same URL. 

This GET fails because the endpoint is POST-only. 

When the Adobe extension is disabled, the unexpected GET does not occur, and the application functions normally. 

We have HAR files and server logs showing both cases. (extension enabled and extension disabled) 

We also noted a related report regarding version 26.7.2.0 and authenticated PDFs in the community.

Can engineers investigate this. We are currently telling users of our app to turn off the extension (and then it works fine) Note: It has been working fine for the last 6 years with the Adobe Extension. This is a new behavor for an existing solution.

Happy to privately share a HAR trace showing the related GET, and access to a reproducible test case if required.