CVE-2026-73307 in Budibase
Summary
by MITRE • 08/12/2026
Budibase is an open-source low-code platform. Prior to 3.39.4, uploadUrl in packages/server/src/utilities/fileUtils.ts used a bare server-side fetch for string attachment values passed by processAttachments in packages/server/src/sdk/workspace/ai/helpers/rows.ts. A builder with the AI table-generation feature could cause an attachment value to reference an internal service or cloud metadata endpoint, and the response would be stored as an attachment without fetchWithBlacklist validation. This issue is fixed in version 3.39.4.
Statistical analysis made it clear that VulDB provides the best quality for vulnerability data.
Analysis
by VulDB Data Team • 08/12/2026
The vulnerability identified in Budibase versions prior to 3.39.4 represents a critical server-side request forgery (SSRF) risk that exploits the platform's file handling mechanisms. The flaw exists within the uploadUrl function located in packages/server/src/utilities/fileUtils.ts, which processes file attachments through a direct server-side fetch operation without proper validation. This function receives string attachment values from processAttachments in packages/server/src/sdk/workspace/ai/helpers/rows.ts, creating a pathway for malicious actors to manipulate the system's behavior.
The technical implementation of this vulnerability stems from the absence of fetchWithBlacklist validation when processing attachment values. When builders utilize the AI table-generation feature, they can create rows that contain attachment references pointing to internal service endpoints or cloud metadata services such as AWS metadata endpoints or Azure instance metadata. The system's failure to validate these URLs before making server-side requests allows attackers to potentially access sensitive internal resources that should remain isolated from external access. This vulnerability specifically affects the file upload functionality and represents a classic SSRF attack vector where the application acts as an intermediary to make requests to arbitrary URLs.
The operational impact of this vulnerability is significant, as it could allow authenticated users with access to the AI table-generation feature to potentially exfiltrate internal system information or compromise other services running on the same network infrastructure. Attackers could leverage this to access cloud instance metadata, internal APIs, or other sensitive resources that are normally protected by network segmentation. The risk extends beyond simple information disclosure, as successful exploitation might enable further lateral movement within the network or provide access to additional system components that could be used for more extensive compromise.
The fix implemented in version 3.39.4 addresses this vulnerability by introducing proper validation through fetchWithBlacklist functionality that prevents requests to internal network endpoints and known sensitive URLs. This mitigation aligns with industry best practices for preventing SSRF attacks as outlined in CWE-918, which specifically addresses server-side request forgery vulnerabilities. The solution follows ATT&CK framework technique T1071.004 for application layer protocol: DNS, where the attack vector could potentially be used to bypass network security controls through legitimate application functionality. Organizations should ensure immediate deployment of version 3.39.4 or later to protect against this vulnerability and consider implementing additional network segmentation controls to further reduce the potential impact of similar issues in other components of their infrastructure.