Summary
A server-side request forgery vulnerability in Budibase REST integrations can allow a builder to redirect an outbound request from a validated public address to an internal address through DNS rebinding. Successful exploitation may provide access to internal services and responses using arbitrary REST methods.
Affected context
The issue occurs in Budibase deployments where REST integrations use the outbound fetch implementation and a user with builder-level permissions can control or influence the target hostname. Although the hostname may initially resolve to an allowed public address, the request path can resolve it again without preserving the validated address, allowing a later DNS response to point to an internal service.
Exploit preconditions
- Attack surface: Budibase REST integration requests
- Required access: Builder-level ability to create or configure an integration
- Trigger point: An outbound REST request after hostname validation
- Network condition: The target hostname must support DNS rebinding or otherwise return a different address during request processing
- Why exploitation is possible: The validation result is not consistently pinned to the connection used by the undici-based dispatcher, creating a time-of-check to time-of-use gap
Why this happens
The underlying bug is a mismatch between address validation and connection establishment. One code path pinned the validated address through a Node agent, while the REST integration used an undici dispatcher that resolved the hostname again and therefore did not honor the original validation result.
This behavior maps primarily to CWE-918: Server-Side Request Forgery, because attacker-influenced destinations can cause the server to access unintended internal resources. It also relates to CWE-367: Time-of-Check Time-of-Use Race Condition, because the address checked during validation may differ from the address used when the connection is established.
Potential impact
- An attacker may access services on internal network interfaces.
- Sensitive responses from cloud metadata or administrative endpoints could be exposed.
- Internal APIs may be invoked using arbitrary REST methods.
- Application credentials or network trust relationships could be abused.
- Compromised internal services may enable further lateral movement.
General mitigation guidance
Mitigation typically focuses on ensuring that destination validation and connection establishment use the same resolved address and trust decision.
Common defensive patterns include:
- Pinning the validated address in the HTTP dispatcher or equivalent connection mechanism.
- Passing validated destination data through custom fetch implementations without re-resolving the hostname.
- Blocking loopback, link-local, private, multicast, and other restricted address ranges for outbound integrations.
- Revalidating redirect destinations and preventing unexpected protocol changes.
- Restricting builder permissions and separating integration configuration from sensitive network access.
- Applying egress network policies and least-privilege service permissions as defense in depth.
How teams detect issues like this
Organizations commonly detect this class of issue through:
- Static analysis for hostname validation followed by a separate DNS resolution or connection step.
- Code review checks for HTTP clients that ignore validated agents, dispatchers, or lookup callbacks.
- Tests using changing DNS answers to verify that the connected address remains the validated address.
- Monitoring outbound requests to private, loopback, link-local, and cloud metadata address ranges.
- Egress policy alerts for unexpected internal destinations or HTTP methods from integration workers.
- Automated security review tools that identify SSRF sinks and time-of-check to time-of-use gaps.

