# GHSA-qr9h-x63w-vqfm: OpenMLS improper persistence of the secret tree during message processing > A bug in OpenMLS fails to update secret tree keys, allowing attackers with local state access to decrypt additional messages, compromising forward secrecy. - **Severity:** Not scored - **Exploited in the wild:** Not known - **Login needed:** Not needed - **Attackable:** Over the internet - **Victim action needed:** Not needed - **Fix available:** Yes - **Published:** 26 Sep 2025 ## What is GHSA-qr9h-x63w-vqfm? GHSA-qr9h-x63w-vqfm is a vulnerability in openmls, published on 26 Sep 2025. An attacker can exploit it over the internet without logging in, with no action from the victim. ### Official description ### Summary A bug in the OpenMLS library prevented private key material from being updated in storage during message processing. The key material in question are the keys stored in the MLS secret tree, which are used for decryption of private MLS messages. The effects of the bug are limited in scope, but can affect forward secrecy and limit how many messages can be decrypted. ### Technical details #### Scope The scope of the bug is limited to private messages (application and handshake messages) received in groups. Furthermore, the scope is limited to one epoch and the effects are reset with each epoch transition, or through consecutive group operations without reloading group state in between. #### Functional impact Within each epoch of a group, there is a maximum number of private messages per sender that can be skipped before an error is thrown. The number of messages is set through maximum_forward_distance in the SenderRatchetConfiguration and the default value is 1000. The bug causes the library to think that private messages have been processed, even when that's not the case. This can cause an error to be returned during message processing. #### Security impact This bug can be exploited only if an adversary gains access to the client’s state. The risk for exploitation is thus relatively low in typical use cases. The bug can be exploited to decrypt more messages than intended by the client. The number of additional messages that can be decrypted is limited to the number of messages that can be skipped, which is maximum_forward_distance. This means that if an adversary gains access to a client’s state, they can decrypt up to maximum_forward_distance additional messages per sender in the current epoch. This affects forward secrecy, as messages that were previously considered secure can be decrypted. OpenMLS otherwise deletes message encryption keys as soon as an epoch ends if not otherwise configured via the `max_past_epochs` setting. This is still the case even in vulnerable versions. ### Affected versions Vulnerable: all versions up to and including version 0.7.0 Not vulnerable: all versions greater than and including version 0.7.1 (see below for security after upgrading from a vulnerable version to a secure one) ### Mitigation Message encryption secrets are deleted when a commit is merged, i.e. when the group transitions to the next epoch. If the client is not configured to retain message secrets for past epochs, messages decrypted in past epochs are thus secure. If past message secrets are retained, messages decrypted in past epochs are secure as soon as the relevant epoch leaves the retention window. Increasing update/commit frequency can thus help keep the window for compromise low. Another way to mitigate the effect is to create a private message (application or handshake message) for each sender of previously received messages. It is important to do that with the same MlsGroup object, withpout reloading it from storage in between. This will ensure that the message encryption keys are updated correctly. ### Security post-patch Since the bug affects persisted state, patching by itself strictly speaking doesn’t solve the issue. State written prior to patching is still vulnerable. An affected client is fully healed from the effects of this bug as soon as it either enters the next epoch (by merging a commit), or, if message encryption key retention is configured, by advancing sufficiently many epochs for pre-patch epoch state to drop out of the retention window. Additionally, the client is healed even before the next epoch, if it processes enough messages s.t. any messages decrypted pre-patch fall out of the window for out-of-order tolerance. This is 5 by default, but can be configured to an arbitrary number using the `sender_ratchet_configuration` setting during group creation. ### Acknowledgements We would like to thank [Ege Erdogan](https://github.com/erdoganege) and [Fatih Ergin](https://github.com/fatihergin) for reporting this issue. ## Which versions are affected by GHSA-qr9h-x63w-vqfm? 2 open source packages ship an affected version, including openmls. | Package | Ecosystem | Affected | Fixed in | | --- | --- | --- | --- | | `openmls` | crates.io | All versions | 0.7.1 | | `openmls` | crates.io | All versions | 0.7.1 | ## How do I fix GHSA-qr9h-x63w-vqfm? Upgrade openmls to 0.7.1 or later. ## Is GHSA-qr9h-x63w-vqfm being exploited? There is no known exploitation of GHSA-qr9h-x63w-vqfm in the wild and no public exploit code. ## Fix prompt for AI coding agents ```text Check this codebase for GHSA-qr9h-x63w-vqfm and fix it if it is affected. What it is: GHSA-qr9h-x63w-vqfm is a vulnerability in openmls, published on 26 Sep 2025. An attacker can exploit it over the internet without logging in, with no action from the victim. Exploitation: There is no known exploitation of GHSA-qr9h-x63w-vqfm in the wild and no public exploit code. Affected packages and fixed versions: - openmls (crates.io): affected all versions; fixed in 0.7.1 - openmls (crates.io): affected all versions; fixed in 0.7.1 Task: 1. Search the dependency manifests and lockfiles (for example package.json, package-lock.json, requirements.txt, poetry.lock, uv.lock, pom.xml, build.gradle, go.mod, Cargo.lock, Gemfile.lock) and any Dockerfiles for the affected packages, including transitive dependencies. 2. If an affected version is in use, upgrade it to the fixed version for its release line with the project's package manager, so the lockfile is regenerated. Prefer the smallest upgrade that fixes it. 3. If there is no fix, or the upgrade would break the build, do not guess: explain the risk and the safest workaround instead. 4. Run the tests and summarise what changed and why. 5. If nothing in the project is affected, say so and change nothing. Reference: https://insights.sbomify.com/GHSA-qr9h-x63w-vqfm/ (Markdown: https://insights.sbomify.com/GHSA-qr9h-x63w-vqfm/index.md) ``` ## Sources - OSV: https://osv.dev/vulnerability/GHSA-qr9h-x63w-vqfm --- From sbomify Insights: https://insights.sbomify.com/GHSA-qr9h-x63w-vqfm/ Summaries are generated and may contain mistakes; the official description above is the record.