Summary
Vikunja contains a broken object-level authorization flaw in its project view deletion API. An authenticated user may reference a view belonging to another project and cause its Kanban assignments and ordering data to be deleted when the request uses an attacker-controlled project identifier.
Affected context
The issue occurs in Vikunja deployments using the affected project view deletion endpoint. It becomes risky when an authenticated user can submit a view identifier independently of the project identifier used for authorization, and the deletion workflow performs follow-up cleanup without confirming that the view belongs to the authorized project.
Exploit preconditions
- Attack surface: Authenticated HTTP API for project view deletion
- Required access: A valid Vikunja user account and permission to access a project
- Trigger point: A request to delete a project view with mismatched project and view identifiers
- Why exploitation is possible: Authorization checks the supplied project but do not establish ownership of the referenced view, while subsequent cleanup uses only the view identifier
Why this happens
The underlying bug is an object authorization mismatch. The application authorizes an attacker-controlled project identifier but fails to verify that the target view is associated with that project, which maps to CWE-639, Authorization Bypass Through User-Controlled Key. The deletion routine also continues after a scoped delete affects no rows and performs related cleanup using only the view identifier, allowing records from another project to be modified or removed.
Potential impact
- An authenticated user could delete task bucket assignments belonging to another project
- An attacker could disrupt Kanban ordering and board presentation
- Cross-tenant integrity violations may occur while the victim view and tasks remain intact
- Users and teams could experience operational disruption or loss of planning metadata
General mitigation guidance
Mitigation generally focuses on enforcing ownership and authorization at
every object boundary in the deletion workflow.
Common defensive patterns include:
- Resolve the target view through the authorized project context rather than trusting independent identifiers
- Verify object ownership before performing the primary deletion or any related cleanup
- Use transactionally consistent, scoped deletes that include both project and view identity
- Treat zero-row or ownership-mismatch results as authorization or not-found failures
- Apply least-privilege database permissions and tenant-aware data access policies
- Add regression tests for cross-project identifier substitution and partial-delete behavior
How teams detect issues like this
Organizations can look for this issue through:
- Static analysis rules identifying delete operations that use user-controlled object identifiers without relationship checks
- Code review signals where authorization uses one path parameter but database cleanup uses another identifier
- Tests that combine valid project identifiers with views from unrelated projects
- API policy checks requiring object ownership validation before destructive actions
- Audit-log monitoring for unusual deletion attempts involving mismatched project and view references
Automated security review tools may help identify missing relationship predicates and authorization checks.

