TGViewer
ThePedroo ThePedroo @thepedr002 · 1.22K subscribers
Post #29 2.45K
Heya, everyone.

Another LSPosed disaster.. Hide My Applist, more widely known by its acronym — HMA — that later got closed source. However, before that happened, it became an essential part of the community as LSPosed usage grew. Then HMA-OSS appeared, fixing detection that targeted the LSPosed's creation rapidly, becoming an appealing alternative to the closed-source module.

But HMA-OSS, unlike the original, was developed towards hiding more properties at the system_server level. Features are fun, but they come with a hefty price; wider surface for detection. Regardless of how careful they were, a detection appeared — and of course, in the *silliest* ways.

With that, I'm here to document some more detections I found regarding HMA-OSS and HMA.

1. Keyboard apps and app list mismatch

Keyboard apps report themselves as one by using BIND_INPUT_METHOD permission in their manifest. This allows the system to recognize such and display them as possible keyboards for the user to select and use.

HMA-OSS introduced a feature that allowed users to hide non-system apps, as some apps have introduced detections against third-party keyboards. However, the applist was not taken in account.

This resulted in a mismatch between InputMethodManager list and the app list, as an application with BIND_INPUT_METHOD permission would not appear as a input method/keyboard app.

2. getPackageInfo timing

Both HMAs intercept getPackageInfo with the purpose of spoofing an app as nonexistent. Hooks, however, are always problematic — well, not directly. Not because they leave traces — which here, they barely do, as we are in system_server, but because the code that you add will always add some overhead to the target function.

By requesting getPackageInfo 20000x, one for HMA(-OSS) app, and one for a nonexistent app, this allowed us to observe (with high-quality performance counters, when HMA is intercepting, a +20% overhead to the function.

OBS: getPackageInfo timing detection was recently discovered by third-parties. However, it was known to us back to December of 2025, and reported to HMA-OSS in January (which is the reason behind it not being detected through it).



I know this wasn't the best one I could make. However, making a post for an almost foreign project, for a detection I made months ago, is hard. My goal here is to inform rather than make a "How To", hence, for this one, no URLs will be provided.

Learn not to be greedy. For the community, help Open Source. Closed-source? Hm..

Sincerely,
ThePedro.

I need a ris- rest.



Go around the clock, hear every tick.
No matter what, it goes on on quick.
Do not be useless, make it be worth.
Leave love behind, leave it for Earth.
  • ❤ 40
  • 🔥 5
  • 👍 1
  • 👏 1
More from @thepedr002
  1. Sep 24, 2026(And ReZygisk 1.0.1, too.Two ones, two two ones!)
  2. Sep 16, 2026(off-topic: Treat Wheel 0.0.11 soon!)
  3. Sep 16, 2026Hello, everyone. Another day, another ignorance. When it comes to RASP (Runtime Applicatio…
  4. Aug 11, 2026(Some credits for me. I've been testing and writing at 2am)
  5. Aug 11, 2026Post #41
  6. Aug 8, 2026Post #40
Threads Profile ViewerView any public Threads profile without an account.Open ThreadLook →Writing with AI? Make it sound human.Metric37 rewrites AI drafts so they read naturally. Free AI detector, 1,500 words free.Try Metric37 →