Summary
Budibase may expose datasource credentials stored in STRING typed fields without redaction, allowing authenticated users with table read permissions to retrieve secrets in plaintext. The exposed values may include MongoDB connection strings and Firebase private keys, potentially enabling unauthorized access to backend services.
Affected context
The issue occurs in Budibase deployments where datasource credentials are stored in STRING typed fields and returned through the read API without adequate redaction. Exploitation requires an authenticated user whose permissions allow access to the affected table or datasource configuration data.
Exploit preconditions
- Attack surface: Budibase read API and datasource configuration records
- Required access: An authenticated account with table read permissions
- Trigger point: Reading records or configuration data containing STRING typed credential fields
- Why exploitation is possible: Credential values are retained and returned in plaintext instead of being consistently masked or excluded from authorized responses
Why this happens
The underlying flaw is insufficient protection of sensitive authentication data during storage and API serialization. Credential-bearing values in STRING typed fields are treated like ordinary application data, allowing response paths to disclose them without applying field-level redaction or secret handling controls. This behavior maps to CWE-522: Insufficiently Protected Credentials.
Potential impact
- Exposure of live MongoDB connection credentials
- Disclosure of Firebase private keys or service account secrets
- Unauthorized access to backend databases or cloud services
- Data theft, modification, or deletion using disclosed credentials
- Further compromise of connected applications and infrastructure
General mitigation guidance
Mitigation generally focuses on reducing the exposure and lifetime of credentials rather than treating secrets as ordinary table data.
Common defensive patterns include:
- Applying consistent server-side redaction and exclusion rules to all credential-bearing fields
- Storing secrets in dedicated secret-management systems instead of general-purpose STRING fields
- Returning opaque references or masked values rather than raw credentials
- Limiting datasource and table permissions according to least privilege
- Rotating credentials that may have been exposed and revoking unnecessary keys
- Preventing sensitive values from appearing in logs, exports, audit records, and API responses
- Using short-lived, narrowly scoped credentials where supported
How teams detect issues like this
Organizations commonly identify this class of issue through:
- Static analysis and code review checks for secret fields serialized without redaction
- API schema and response tests that assert credential-bearing properties are omitted or masked
- Inventory scans for connection strings, private keys, and tokens in database records
- Permission reviews confirming that table read access does not implicitly grant secret access
- Runtime monitoring for unusual reads of datasource configuration records
- Policy enforcement that blocks storage of raw credentials in general-purpose application fields
Automated security review tools may help detect plaintext secret handling and missing field-level access controls.

