GHSA-38f9-m297-6q9g is a medium-severity (CVSS 5.9) SQL Injection vulnerability in @nozbe/watermelondb. O3 Security confirms whether GHSA-38f9-m297-6q9g is actually reachable in your code before you act, and blocks exploitation at runtime until you patch.
DoS via malicious record IDs in WatermelonDB
Real-World Exposure
How broadly this vulnerability is actually deployed: weekly install volume shows current usage, and reverse-dependency count shows how many other packages break if it stays unpatched.
@nozbe/watermelondbnpmDescription
Impact
Medium severity 5.9 https://www.first.org/cvss/calculator/3.0#CVSS:3.0/AV:N/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:H
A maliciously crafted record ID can exploit a SQL Injection vulnerability in iOS adapter implementation and cause the app to delete all or selected records from the database, generally causing the app to become unusable.
This may happen in apps that don't validate IDs (valid IDs are /^[a-zA-Z0-9_-.]+$/) and use Watermelon Sync or low-level database.adapter.destroyDeletedRecords method.
The integrity risk is low due to the fact that maliciously deleted records won't synchronize, so logout-login will restore all data, although some local changes may be lost if the malicious deletion causes the sync process to fail to proceed to push stage.
No way to breach confidentiality with this vulnerability is known. Full exploitation of SQL Injection is mitigated, because it's not possible to nest an insert/update query inside a delete query in SQLite, and it's not possible to pass a semicolon-separated second query. There's also no known practicable way to breach confidentiality by selectively deleting records, because those records will not be synchronized.
It's theoretically possible that selective record deletion could cause an app to behave insecurely if lack of a record is used to make security decisions by the app.
Patches
Patched versions include:
- 0.15.1
- 0.16.2
- 0.16.1-fix
- this is actually the same as 0.16.0, but with the patch applied - as 0.16.1 is causing issues for some users
924c7ae2a8dcommit id contains the patch
Workarounds
- Ensure that your backend service sanitizes record IDs sent in the
pull syncendpoint, such that only IDs matching/^[a-zA-Z0-9_-.]+$/are returned. This could also be done in JavaScriptpullChangesfunction passed tosynchronize() - If you use
destroyDeletedRecordsdirectly, validate all IDs passed the same way
For more information
If you have any questions about this advisory, contact @radex.
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 📦npm | @nozbe/watermelondb | all versions | 0.15.1 |
| 📦npm | @nozbe/watermelondb | ≥ 0.16.0&&< 0.16.2 | 0.16.2 |
Detection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for @nozbe/watermelondb. O3's reachability analysis confirms whether the vulnerable code path is actually invoked in your application, so you act on real exposure instead of every transitive match.
Fix
Update @nozbe/watermelondb to 0.15.1 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-38f9-m297-6q9g is resolved across your whole dependency graph.
Workarounds
If you can't upgrade right away: gate or disable the affected feature, validate untrusted input at the boundary, and avoid passing attacker-controlled data into the vulnerable path. O3's runtime protection blocks exploitation in production as an interim safeguard until the upgrade lands.
How O3 protects you
O3 pinpoints whether GHSA-38f9-m297-6q9g is reachable in your code and exactly where to fix it, then blocks exploitation in production at runtime until the patched version is deployed.
Tailored to GHSA-38f9-m297-6q9g. Runtime protection reduces exposure until a permanent patch is applied and verified — it complements patching, it doesn't replace it.
Frequently Asked Questions
Is GHSA-38f9-m297-6q9g in your dependencies?
O3 detects GHSA-38f9-m297-6q9g across npm dependencies and uses function-level reachability to confirm whether the vulnerable code path is actually reachable — not just present. No false positives.