Splitting real installs from Cache Placeholder "Installs"

[McAwesome]

My organization has opted to index the /Users/ directory for various reasons.  This hasn't been a big deal until I got a request to patch an application where the dev reused their app name and bundleID on the macOS and iOS versions.  As a result, searching for either the Application Name or BundleID catches machines with it in /Applications/ and machines that have a placeholder in ~/Library/Daemon Containers/<device info>/Data/Library/Caches/Placeholders-v2.noindex. 

I'm kinda stumped on the best way to scope a smart group to include installs in /Applications/ or ~/Applications but exclude that placeholder directory.  Usually, the devs have slightly different bundle IDs we can use to make things more targeted.

Does anyone here have any recommendations for the best way to scope a group so that it doesn't catch those placeholders locations?

[AJPinto]

This sounds like an Application Development problem not a MDM problem. You could write an Extension Attribute to read a specific value you can use to distinguish, but it wont report the same as the rest of your applications.

[McAwesome]

Agreed, but knowing the app developer I'm not sure they'll bother fixing things like this.  I've opened a ticket with Jamf about the False Positive detection on their Patch Reporting title.  We'll see where that goes.

[AJPinto]

Honestly, this isn’t something Jamf can or should fix. Their patch reporting relies on Apple’s documented best practices, which nearly every app developer follows (Adobe being a rare exception). The issue is that your internal devs decided to ignore those practices, and that’s what’s causing the false positives.

If this reporting is important, I’d suggest bringing it up internally. Reusing bundle IDs across platforms and dropping app-related files in nonstandard locations isn’t something Jamf can work around. Expecting them to build in support for a one-off internal app that doesn’t follow platform conventions isn’t realistic. This is something that really needs to be fixed at the source.

(Edited my response making it more professional)

[McAwesome]

The application in question is ChatGPT, not something my organization develops.

(Edit) I should also add that I can see benefit in Jamf adding some safety checks so that items in a pending sync state to or in a backup of an iOS device do not get inventoried the same way as traditionally installed applications. This is the core of the issue I'm having.  There's a standard macOS managed directory in the user's Library for all that, so it is something that could be excluded.  The other upside to doing that is it'd make any searches or groups relying on application name instead of bundle ID would become more accurate as well.  

The reused bundleID is just an extra wrinkle OpenAI threw into things that complicates addressing it from my side of things.  I'd love to live in a world where we could assume everyone is doing Best Practices on things like that, but we have to deal with the reality that there are simply too many developers out there for that assumption to work at any major scale.