Summary
Budibase's REST datasource query preview can send server-side requests to destinations in the shared address space 100.64.0.0/10 because that range is missing from the default SSRF blacklist. An authenticated user with Builder permission may use this behavior to reach an accessible HTTP(S) service and receive its response through the preview flow when the deployment relies on the default blacklist.
Affected context
The issue occurs in Budibase self-hosted deployments using the backend-core REST datasource query preview functionality. It is relevant when no custom BLACKLIST_IPS value is defined, leaving the default blacklist active without an entry for 100.64.0.0/10, and when a service in that range is reachable from the Budibase server.
Exploit preconditions
- Attack surface: REST datasource query preview functionality exposed through the Budibase server
- Required access: An authenticated account with Builder permission
- Trigger point: A REST datasource query preview request processed by the backend
- Network condition: The target HTTP(S) service in
100.64.0.0/10must be reachable from the server - Why exploitation is possible: The default destination blacklist does not cover the shared address space, allowing the server to make and relay requests that should have been blocked
Why this happens
The vulnerability is caused by incomplete SSRF destination filtering. The default blacklist used for REST datasource previews omits the shared address space range 100.64.0.0/10, so a server-side request can be directed to an otherwise restricted network location when that range is reachable.
This behavior maps to CWE-918: Server-Side Request Forgery, in which user-influenced request destinations are insufficiently constrained and the application acts as a proxy from a more trusted network position.
Potential impact
- An attacker may access reachable internal or carrier-grade network services
- Sensitive response data may be disclosed through the preview workflow
- Internal APIs or administrative endpoints could be exposed if reachable
- Requests could cause unintended actions on services that trust the Budibase server
- The vulnerability could support reconnaissance and follow-on attacks against accessible infrastructure
General mitigation guidance
Mitigation generally focuses on enforcing positive destination controls rather than relying only on a partial blacklist.
Common defensive patterns include:
- Add
100.64.0.0/10to the effective SSRF deny policy when that range is not intended as a valid destination - Prefer allowlists of approved schemes, hosts, ports, and network destinations for datasource previews
- Resolve and validate destinations consistently, including redirects and alternate address representations
- Apply outbound network controls that prevent application containers from reaching sensitive or unnecessary network segments
- Restrict query preview capabilities to trusted roles and minimize the permissions of the Budibase service account
- Avoid returning full upstream responses when preview functionality does not require them
How teams detect issues like this
Detection can combine application, network, and source-level controls:
- Review REST datasource preview requests and responses for destinations in
100.64.0.0/10or other private and special-use ranges - Monitor outbound connections from Budibase services to unexpected internal, shared, or administrative addresses
- Use static analysis and code review rules to identify user-controlled URLs passed to HTTP clients without centralized destination validation
- Check configuration and deployment policy to verify that effective SSRF controls include all prohibited address ranges
- Enforce network egress policies and alert on preview traffic that crosses approved service boundaries
- Automated security review tools may help identify incomplete blacklist logic and missing SSRF allowlist enforcement

