Summary
A critical authorization bypass in LXD allows an authenticated attacker to evade project security restrictions during cross-project instance migrations. When the migration is directed to another cluster member, the destination may treat the request as trusted internal traffic and accept instance configurations that the target project should prohibit.
Affected context
This issue may occur in LXD clusters that permit authenticated users to initiate cross-project instance migrations. The risk is present when a migration request identifies both a target project and a different cluster member, while the destination node processes the request as an internal cluster notification and therefore does not repeat the applicable project restriction checks.
Exploit preconditions
- Attack surface: LXD's authenticated instance migration API
- Required access: Authenticated permission to initiate or otherwise control an instance migration
- Trigger point: A cross-project migration directed to a different cluster member
- Why exploitation is possible: The destination node may trust the migration as an internal notification and omit authorization checks that should enforce restrictions for the target project
Why this happens
The underlying flaw is an authorization boundary failure in distributed request handling. LXD applies project restriction checks inconsistently between the originating and destination nodes, allowing a request that crosses a trust boundary to be treated as internally authorized without revalidating the target project's policy. This behavior maps to CWE-863: Incorrect Authorization, because access control decisions are not enforced for the requested operation and resource context.
Potential impact
- An attacker may introduce prohibited instance configurations into a restricted project.
- Project-level isolation controls could be weakened.
- Unauthorized compute, storage, or network resources could be consumed.
- Affected projects could experience policy violations or operational disruption.
- The bypass could support further attacks if the introduced configuration has elevated capabilities.
General mitigation guidance
Mitigation generally focuses on enforcing authorization consistently at every node and trust boundary.
Common defensive patterns include:
- Revalidate the caller's permissions and target-project restrictions on the destination node.
- Treat migration metadata as untrusted until authenticated and authorized independently.
- Use secure-by-default project policies that deny prohibited configurations unless explicitly permitted.
- Separate internal transport authentication from authorization to access project resources.
- Apply least privilege to migration roles and limit cross-project migration capabilities.
- Maintain auditable, tamper-resistant records of migration requests and authorization decisions.
How teams detect issues like this
Organizations can look for this class of issue through:
- Static analysis of distributed request handlers for authorization checks that occur only on the originating node.
- Code review signals showing internal-notification paths that bypass normal policy or project validation.
- Tests that exercise cross-project operations across different cluster members and compare decisions at both ends.
- Policy enforcement requiring explicit authorization for every destination project and resource type.
- Audit monitoring for authenticated users creating instances or configurations in projects outside their normal scope.
- Automated security review tools that flag trust-boundary transitions without repeated authorization.

