LSPosed-Irena: a Zygisk fork that ships upstream's README and calls itself useless
Useless LSPosed Framework Fork
At a glance
- What is it?
- LSPosed-Irena is a fork of the LSPosed ART hooking framework that runs as a Zygisk module and is distributed through Telegram, with canary builds coming out of GitHub Actions. Its documentation is upstream's own, so nothing in it says what the fork changed.
- Who is it for?
- Treat this repository as a build of LSPosed rather than a documented product: if you are a module developer, the parts worth reading are the supported range, the two distribution channels and the statement that modules are API compatible in both directions, and everything else you will find is upstream text.
- Can I use it commercially?
- Yes, with conditions. GPL-3.0 is a copyleft licence: if you distribute software that includes it, you must release that software's source code under the same licence. Running it internally without distributing it does not trigger that obligation.
- Is it still maintained?
- Yes. The repository last received commits 177 days ago.
- What is it written in?
- Mainly Java, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.
Star history
- re-zero001/LSPosed-Irena2,680 ★ · +1
| Repository | 2026-10-04 | 2026-11-15 |
|---|---|---|
| re-zero001/LSPosed-Irena | 2.7k | 2.7k |
1 Stars in the last 42 days, measured from the daily GitHub snapshot. Compare with other projects
Editorial analysis
The pages are upstream LSPosed's own, so the fork's changes are undocumented
The repository description reads Useless LSPosed Framework Fork, and the README is titled LSPosed Framework with the upstream introduction, the upstream explanation of what Xposed is for, and the upstream credits. The fork relationship is stated in the credits rather than in a description of changes: LSPosed is credited as the fork source and EdXposed as an LSPosed fork source, so this is a fork of a fork. The Community Discussion section says only that this fork won't have one.
What that leaves is a documentation gap rather than a contradiction. There is no section anywhere in these pages that says what Irena changes relative to upstream, no list of behavioural differences, and no note on why a second fork exists when the original is still linked in the header badges next to this repository's own releases link. Anyone evaluating the fork has to read commits on the dev branch to find out what moved, and the README does not point them there.
The Dobby entry adds a small hint. Inline hooking is credited to Dobby, and the link goes to a Dobby fork under the same author rather than to the upstream Dobby project, so at least one dependency is a personal fork too. That is the kind of detail a changelog should carry. Here it sits in a credit list as a bare link.
Android 8.1 to 16 is the entire support statement
Supported Versions is one line: Android 8.1 ~ 16. There is no per version note, no list of known broken releases and no statement about which Android versions the modules built against it are expected to work on.
The install path sets the floor instead. It asks for Magisk v26+, then ZygiskNext, with a note that you can enable zygisk in Magisk instead but that ZygiskNext is what the authors recommend, then installing LSPosed in the Magisk app, rebooting, and opening the manager from the notification, with an alternative of installing manager.apk from the zip file. The last step of that list is Have fun.
That sequence is a rooted device with a specific Magisk generation, a second Zygisk implementation, and a reboot, on a framework whose stated job is changing the behaviour of the system and of apps without touching any APK. The version range is generous and the installation requirements are not, and both facts belong in the same place when you are deciding whether a given phone can carry this.
Injection into zygote has no per app boundary, no dry run and no off switch
The credits state what ZygiskNext provides: a way to inject code into the zygote process. Zygote is the process every app forks from, so code injected there is present in every application the device launches, which is the mechanism behind the claim in the introduction that modules can change the behaviour of the system and apps without touching any APKs.
Now put that next to what these pages do not contain. There is no per application enable list, no dry run mode, no staged rollout and no abort switch described for the framework itself. The only undo statement in the README is the upstream one about Xposed modules: all changes are done in memory, so you deactivate the module and reboot to get your original system back. That sentence is about a module, not about a framework that installs itself as a Zygisk module, adds its own manager app and injects into zygote.
For an engineer this is the difference between a debugging tool and a system change. There is no scope to review, so the question of what a given module is allowed to touch has to be answered outside the framework, before installation.
Distribution runs through Telegram and Actions, and only canaries get bug reports
The Download section has two entries. The first is a Telegram channel, @lsposed-irena, linked through its t.me share form. The second is a link to the core GitHub Actions workflow filtered on branch dev, described as where to check for canary builds. Neither of those is a release. The tagged releases on the releases page are 2.0.0 from 2026-05-14 and 1.9.2 from 2024-10-17, so the two most recent tags are about nineteen months apart, while the last push to the default branch, dev, was on 2026-05-22.
The support policy then rules out the tags. The Get Help section accepts only bug reports from the latest debug build, and issues are filed on GitHub, with a separate note that the project accepts English issue titles only and suggests a translation tool for Chinese speakers. So the builds you can get from the releases page are the builds you cannot report problems about, and the build you should report problems about is whatever the most recent workflow run on dev produced.
For a framework other people's modules depend on, that ordering is worth stating plainly before you install it, because it decides where you look for a known good build and it decides where you will be sent when something breaks.
The module repository sentence says our own and links upstream
The For Developers section invites module authors and claims compatibility in both directions: a module based on the LSPosed framework is fully compatible with the original Xposed framework, and a module written against the original works well here too. The Xposed Framework API link points at api.xposed.info, which is upstream documentation rather than anything written for this fork.
Then comes the sentence that does not add up. It says we use our own module repository and that developers are welcome to submit modules to our repository, so modules can then be downloaded in LSPosed. The link on our repository goes to github.com/Xposed-Modules-Repo, which is the module index shared with the upstream framework, not a repository hosted under this fork's owner.
So the submission target for module authors is the shared index, while the surrounding sentence describes it as a private one. That is a small thing for a user and a real one for a module author, because where you submit decides who sees your module and which framework's manager lists it.
Four struck through credits record the move away from YAHFA to LSPlant
The credits read partly as a change log for the hooking strategy. The active entries are LSPlant as the core ART hooking framework, Dobby for inline hooking, XposedBridge for the original Xposed APIs, xz_embedded for decompressing the debug_info section into stripped libraries, and system_properties for switching properties access within LSPosed.
Four more entries are struck through with strikethrough markers: SandHook as an ART hooking framework for the SandHook variant, YAHFA as the previous ART hooking framework, dexmaker and dalvikdx for dynamically generating YAHFA hooker classes, and DexBuilder for the same purpose. Every struck entry concerns YAHFA or a YAHFA variant. What that tells you is that the previous approach generated hooker classes at runtime for a YAHFA style hook, and that approach has been replaced rather than kept alongside.
Two of the active credits are also useful as tells about what this framework has to do at runtime. Decompressing debug_info into stripped libraries means reading symbols that release builds removed, and switching properties access means altering how system properties behave inside the process. Neither is a detail a module author needs to configure, and both are consequences of running inside someone else's process.
The tree is a full application stack with a bundled dex2oat
The root of the repository holds a Gradle project rather than a patch. There is build.gradle.kts, settings.gradle.kts, gradle.properties, a gradle directory, gradlew and gradlew.bat, plus a .gitmodules file, which is how a fork pulls in the projects it builds from. The source side is split into app, core, daemon, services, magisk-loader, dex2oat, hiddenapi, libxposed and external.
Two of those names carry weight. A magisk-loader directory is the piece that presents itself inside the Magisk app for the install step, and a dex2oat directory is the ART ahead of time compiler, which is a substantial native binary rather than a wrapper script. hiddenapi and libxposed point at the modern API surface modules are expected to target, which is consistent with the claim that modules work in both directions.
The build also carries crowdin.yml, so translations run through Crowdin, and .gitattributes and .gitignore at the root. Nothing about this layout tells you how far the fork has diverged from upstream, since a directory tree shows structure and not behaviour, but it does show that a build is a real part of the project and that a fork which ships no documented diff is asking you to read its tree instead.
Nothing on these pages says how to take the framework back off a device
There is no uninstall section. The install list runs Magisk, ZygiskNext, install in the Magisk app, reboot, open the manager from the notification, and stops there. The only removal-adjacent statement in the whole README is the upstream paragraph about Xposed modules, which says that because all changes are done in memory, deactivating the module and rebooting gives you your original system back.
That claim is about a module's effect on the code it hooked, not about the framework's own footprint. A Zygisk module that injects into zygote, ships a manager application and hooks the ART runtime is not undone by deactivating one of its own modules, and the pages here do not say what it takes to remove it from a device.
This is the gap to close before installing, not after. If the device holds anything you cannot restore from a backup, that missing procedure is a reason to look for the upstream LSPosed documentation, which is linked from the badges in the header.
Editorial conclusion
Treat this repository as a build of LSPosed rather than a documented product: if you are a module developer, the parts worth reading are the supported range, the two distribution channels and the statement that modules are API compatible in both directions, and everything else you will find is upstream text. Before you rely on it, check three things the README does not answer, namely whether the build you installed came from Telegram or from a workflow run, whether a tagged release corresponds to what you are running, and how you would remove it, because no uninstall path is written down anywhere in these pages. Only bug reports against the latest debug build are accepted, so a tagged release is the worst thing to file an issue against. Installing any of this needs Magisk v26+, ZygiskNext and a reboot on a rooted device, which makes it a poor fit for a phone carrying daily credentials. Licence is GPL-3, and the README names upstream LSPosed as the fork source.
Frequently asked questions
What does the LSPosed framework do?
It is a Zygisk module that provides an ART hooking framework, delivering APIs consistent with the original Xposed so modules can change the behaviour of the system and of apps without modifying any APK. Changes are made in memory, and it hooks ART through the LSPlant framework.
What is the difference between Xposed and LSPosed?
The README frames Xposed as the original framework and says this project delivers consistent APIs with it while leveraging LSPlant. It states that modules are compatible in both directions, without measuring one against the other.
How do you remove the LSPosed framework from a phone?
These pages document no uninstall procedure. The only related statement is the upstream note that a module's in memory changes are undone by deactivating the module and rebooting, which does not cover the framework's own Zygisk module and manager app.
Is the LSPosed-Irena fork safe to install on a daily phone?
The README does not assess risk. What it states is the mechanism and the cost: ZygiskNext injects code into the zygote process, installation needs Magisk v26+, ZygiskNext and a reboot, and there is no per application enable list, dry run or off switch documented for the framework itself.
Where do I download an LSPosed-Irena build?
Two channels are given: the Telegram channel @lsposed-irena, and the core GitHub Actions workflow on branch dev for canary builds. The tagged releases are 2.0.0 from 2026-05-14 and 1.9.2 from 2024-10-17, and only bug reports from the latest debug build are accepted.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/re-zero001-lsposed-irena)Star history chart for your README
[](https://hysenlabs.com/projects/re-zero001-lsposed-irena)Was this useful?