Vulnerability call chain
1.1 Summary
DSpace has a missing authorization vulnerability: EPerson byEmail Search Leaks Account Authorization Properties. Unauthorized disclosure of EPerson account attributes including email, netid, canLogIn, requireCertificate, selfRegistered, and lastActive. These are authorization/authentication-related properties, especially netid and login flags
- Attack precondition: The attacker can call
GET /api/eperson/epersons/search/byEmailand knows or guesses a target email address - Security impact: Unauthorized disclosure of EPerson account attributes including
email,netid,canLogIn,requireCertificate,selfRegistered, andlastActive. These are authorization/authentication-related properties, especiallynetidand login flags
1.2 Exploit path
GET /api/eperson/epersons/search/byEmail?email=<target> calls findByEmail() without method-level @PreAuthorize. The result is converted to EPersonRest, exposing account fields that normally require EPERSON READ via the direct findOne endpoint
1.3 Key code evidence
EPersonRestRepository.java
Evidence location: EPersonRestRepository.java
2. EPersonConverter.java
Evidence location: EPersonConverter.java
3. EPersonRestPermissionEvaluatorPlugin.java
Evidence location: EPersonRestPermissionEvaluatorPlugin.java
3. Root Cause Analysis
Root Cause 1: Missing server-side authorization on the vulnerable operation.
The endpoint accepts user-controlled authorization-sensitive identifiers or fields, but the write/read path does not prove that the current caller may operate on the target object.
Root Cause 2: Missing object-scope or grant-bound validation.
The implementation relies on endpoint access, UI filtering, or object existence checks instead of enforcing target ownership, tenant boundary, role ceiling, or grantable-resource constraints at the service layer.
4. Recommended fix
Add @PreAuthorize("hasPermission(#email, 'EPERSON', 'READ')") equivalent logic after resolving the target EPerson, or restrict byEmail to admin / self and return a minimized response for other users
5. Verification after fix
- Unauthorized callers receive HTTP 403 or equivalent rejection.
- Out-of-scope target identifiers are rejected before database writes or sensitive reads.
- Role, permission, tenant, organization, ownership, or grant-bound ceilings are enforced server-side.
- Direct HTTP requests are rejected even when front-end controls are hidden.
