GHSA-j8v8-g9cx-5qf4
HIGHGHSA-j8v8-g9cx-5qf4 is a high-severity (CVSS 8.3) CWE-639 vulnerability in @better-auth/scim. O3 Security confirms whether GHSA-j8v8-g9cx-5qf4 is actually reachable in your code before you act, and blocks exploitation at runtime until you patch.
@better-auth/scim: Account/provider takeover via missing owner binding on non-org SCIM providers
Real-World Exposure
How broadly this vulnerability is actually deployed: weekly install volume shows current usage, a proxy for how much of the ecosystem is exposed.
@better-auth/scimnpmDescription
Am I affected?
Users are affected if all of these hold:
- They install and register the
@better-auth/scimplugin (plugins: [scim()]). - They create SCIM providers without an
organizationId, that is, non-organization ("personal") providers. Organization-scoped providers are not affected because they enforce organization membership and role. - Their application's deployment has more than one authenticated user, so one account can target another user's provider.
- They have not set
providerOwnership: { enabled: true }.
The SCIM management endpoints first shipped in 1.5.0, so every stable release from 1.5.0 onward is affected, and the stable line is not patched. On the pre-release line, builds through 1.7.0-beta.3 are affected, and 1.7.0-beta.4 carries the fix.
Fix:
- Upgrade to
@better-auth/[email protected](then1.7.0). This is a breaking change: it removes theproviderOwnershipoption, makes owner binding mandatory, and adds a permanentscimProvider.userIdcolumn. - Run the schema migration after upgrading (
npx auth migrate). Connections created before the upgrade carry no owner and become unreachable through the management endpoints, so reclaim them at the database level. - The
1.6.xstable line is not patched. If developers stay on it, apply the workaround below.
Summary
@better-auth/scim does not bind non-organization SCIM providers to their creator in the default configuration. Any authenticated user can manage another user's non-org provider, including reading its metadata, listing connections, regenerating its SCIM bearer token, and deleting the connection. Regenerating the token rotates it: the legitimate token stops working and the attacker holds a valid one.
Details
The plugin tracks provider ownership through an opt-in providerOwnership option and a scimProvider.userId column. Both default to off. When ownership is disabled, a non-org provider row is created without a userId, and the management access check denies access only when a stored owner id is present and differs from the caller:
// non-organization branch of the access check
} else if (provider.userId && provider.userId !== userId) {
throw new APIError("FORBIDDEN", { message: "You must be the owner to access this provider" });
}
Because the default row has no userId, the condition is false and the request is allowed. The same gap reaches every management operation: token generation and regeneration, single-provider read, the connection list (which returns providers where p.userId === userId || !p.userId), and provider deletion.
The most sensitive operation is token regeneration. If Alice creates a non-org provider corp-idp, Bob can call the same generate-token endpoint with providerId: "corp-idp". The existing row passes the ownerless access check, so the plugin deletes it, creates a fresh row, and returns a new SCIM bearer token to Bob. Alice's previous token is now invalid, and Bob's token authenticates against that provider's SCIM API routes.
This requires no organization membership and no elevated role. It is reachable in the default scim() configuration whenever non-organization providers are used.
Proof of concept
Two authenticated users, Alice and Bob, in a default scim() setup:
- Alice signs in and creates a non-org provider:
POST /scim/generate-tokenwith{ "providerId": "corp-idp" }. - Bob signs in and reads Alice's provider:
GET /scim/get-provider-connection?providerId=corp-idpreturns200with{ "providerId": "corp-idp" }. - Bob regenerates the token:
POST /scim/generate-tokenwith{ "providerId": "corp-idp" }returns201and a newscimToken. - Alice's old token is rejected at
GET /scim/v2/Users(401); Bob's regenerated token is accepted (200).
The behavior is reproducible through Better Auth's public API surface with two distinct sessions.
Patches
Fixed in @better-auth/[email protected] (then 1.7.0). The fix makes owner binding mandatory: generateSCIMToken records the creator's userId on every personal connection, and the management endpoints grant access only to that owner. Organization-scoped connections keep their existing membership and role checks. The 1.6.x stable line is not patched.
This release is breaking. It removes the providerOwnership option, owner binding can no longer be disabled, and the scimProvider.userId column becomes a permanent part of the schema, so run npx auth migrate after upgrading. Connections created before the upgrade carry no owner and fail closed; reclaim them by deleting scimProvider rows that have neither organizationId nor userId, or by setting userId to the intended owner, then regenerating tokens.
Workarounds
Until a patched version is available:
- Set
providerOwnership: { enabled: true }when registering the plugin, then run the schema update so thescimProvider.userIdcolumn exists (npx auth generateornpx auth migrate). New non-org providers are then owner-bound and non-owners are denied. Providers created before enabling ownership stay ownerless until recreated. - Or scope every SCIM provider to an organization by always passing
organizationId. Organization providers enforce membership and role and are not exposed. - Or restrict access to the SCIM management endpoints at the edge while non-org providers remain ownerless.
Impact
Any authenticated user can take over non-organization SCIM providers created by other users when the plugin runs with its default ownership configuration. An attacker can discover non-org provider connections, read provider metadata, regenerate another user's SCIM token (invalidating the legitimate one), authenticate to the SCIM API routes with the new token, manage SCIM-provisioned users subject to the enabled SCIM functionality, and delete the provider connection.
Credit
Reported by Jvr2022 through a private GitHub Security Advisory.
Resources
- CWE-862 Missing Authorization: https://cwe.mitre.org/data/definitions/862.html
- CWE-639 Authorization Bypass Through User-Controlled Key: https://cwe.mitre.org/data/definitions/639.html
- SCIM 2.0 Protocol (RFC 7644): https://datatracker.ietf.org/doc/html/rfc7644
- OAuth 2.0 Bearer Token Usage (RFC 6750): https://datatracker.ietf.org/doc/html/rfc6750
Affected Packages
| Ecosystem | Package | Vulnerable range | Fix |
|---|---|---|---|
| 📦npm | @better-auth/scim | ≥ 1.5.0&&< 1.7.0-beta.4 | 1.7.0-beta.4 |
Detection & mitigation playbook
Open-source dependencyDetect
Scan your dependency tree (package-lock.json, pnpm-lock.yaml, requirements.txt, go.sum, etc.) for @better-auth/scim. 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 @better-auth/scim to 1.7.0-beta.4 or later, then make sure no transitive (indirect) dependency still pins the vulnerable range — O3 confirms GHSA-j8v8-g9cx-5qf4 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-j8v8-g9cx-5qf4 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-j8v8-g9cx-5qf4. 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-j8v8-g9cx-5qf4 in your dependencies?
O3 detects GHSA-j8v8-g9cx-5qf4 across npm dependencies and uses function-level reachability to confirm whether the vulnerable code path is actually reachable — not just present. No false positives.