Summary
Budibase exposes an unauthenticated endpoint that generates S3 PutObject presigned URLs using workspace datasource credentials.
An attacker who knows a workspace ID and S3 datasource ID can obtain signed upload URLs for arbitrary buckets and keys, creating a high-risk path for unauthorized storage access and data manipulation.
Affected context
This vulnerability occurs in Budibase application server deployments where:
- S3 datasources are configured within workspaces
- The presigned URL generation route is reachable from the public network
- recaptcha middleware is used as the only protection for the endpoint
Under these conditions, the service may issue presigned S3 PutObject URLs without enforcing standard authentication or authorization checks on the caller.
Exploit preconditions
- Attack surface: Public HTTP endpoint on the Budibase application server
- Required access: Knowledge of a target workspace ID and associated S3 datasource ID
- Trigger point: Calling the presigned URL generation route with attacker-chosen bucket and key parameters
- Why exploitation is possible: The endpoint relies solely on recaptcha middleware and does not enforce authentication, workspace membership, table permissions, or datasource access controls, allowing untrusted callers to leverage internal S3 credentials.
Why this happens
The underlying issue maps to CWE-306: Missing Authentication for Critical Function.
A server-side route that performs a sensitive operation—creating S3 PutObject presigned URLs backed by trusted workspace credentials—is exposed without proper authentication or authorization gates. By treating presigned URL generation as a low-risk, captcha-protected action rather than a privileged function, the design implicitly grants storage capabilities to any caller who can reach the endpoint and provide minimal identifiers. This insecure pattern results from misclassifying the security sensitivity of helper APIs and failing to consistently apply identity and permission checks to data access flows.
Potential impact
- Unauthorized uploads to S3 buckets using workspace-bound credentials
- Data pollution or overwriting of existing S3 objects where keys collide
- Abuse of organization-managed storage for hosting arbitrary attacker content
- Indirect compromise of application data integrity if S3-backed resources are consumed by Budibase apps
- Potential stepping stone for further attacks that rely on trusted storage artifacts
General mitigation guidance
Effective mitigation focuses on treating presigned URL generation as a privileged capability and enforcing strict identity and access controls around it.
Useful patterns include:
- Requiring authenticated user context and validating workspace membership before generating presigned URLs
- Applying fine-grained authorization checks tied to datasource configuration and table-level permissions
- Limiting which buckets, prefixes, and keys can be targeted based on trusted server-side configuration rather than user input
- Centralizing access control logic so that all data-accessing routes consistently enforce authentication and permission policies
- Reducing reliance on soft protections (like captchas) for operations that expose infrastructure credentials or powerful data primitives
How teams detect issues like this
Detection commonly relies on a mix of design review and technical controls:
- Reviewing API specifications and server routes for endpoints performing sensitive actions without authentication or authorization checks
- Static analysis or policy-as-code rules that flag handlers generating presigned URLs or accessing cloud credentials without preceding auth middleware
- Logging and monitoring access to presigned URL generation endpoints, looking for anonymous or unusually frequent calls from untrusted origins
- Continuous security testing of low-code platform APIs, including checks for undocumented or weakly protected helper endpoints

