CVE-2025-36324: IBM watsonx.data Intelligence SSRF Vulnerability - Patch & Detection Guide

Technical summary

CVE-2025-36324 is classified as CWE-918, a server-side request forgery vulnerability in IBM watsonx.data intelligence. The flaw allows authenticated attackers to craft malicious requests that the affected system will execute on their behalf, potentially accessing internal services, cloud metadata endpoints, or other network resources not directly accessible from the attacker's position. The vulnerability requires valid authentication credentials to exploit, meaning internal users or compromised accounts pose the primary risk vector.

Business impact

This vulnerability increases insider threat and account compromise risks. If an internal user is compromised or acts maliciously, attackers gain a trusted launch point for network reconnaissance and lateral movement. Organizations using watsonx.data intelligence for analytics workloads may face disruption if the system is leveraged to attack downstream infrastructure. The SSRF can also facilitate information disclosure about internal network topology and service availability, undermining network segmentation strategies.

Affected systems

IBM watsonx.data intelligence versions 5.2.0, 5.2.1, 5.2.2, and 5.3.0 are vulnerable. IBM Software Hub, which may bundle or depend on watsonx.data intelligence, may also be affected. Organizations should identify all deployments of these versions in their environment and prioritize patching or isolation of affected instances.

Exploitability

Exploitation requires valid authentication credentials, which moderates the attack surface compared to unauthenticated SSRF flaws. However, the CVSS score of 4.3 (MEDIUM) reflects that once an attacker is inside the authentication boundary—whether through credential compromise, insider threat, or legitimate user action—the SSRF is relatively straightforward to trigger. No user interaction is required beyond sending a crafted request. This is not currently tracked in CISA's Known Exploited Vulnerabilities (KEV) catalog.

Remediation

Patch all affected instances to versions newer than 5.3.0, following IBM's official security advisories and release notes. Verify the specific patch version number against IBM's vendor advisory before deployment. In the interim, restrict network access from watsonx.data intelligence systems to only required external services, enforce network segmentation to limit lateral movement if SSRF is exploited, and audit access logs for suspicious outbound connections originating from the system.

Patch guidance

Check IBM's security bulletin for watsonx.data intelligence to identify the minimum patched version. Deploy patches in a test environment first to validate compatibility with existing analytics workflows and integrations. For air-gapped or offline deployments, verify that patch availability and delivery mechanisms align with your infrastructure. Consider scheduling patching during maintenance windows to minimize disruption to data intelligence operations. Test outbound network connectivity after patching to ensure business functions remain intact.

Detection guidance

Monitor for unusual outbound connections from watsonx.data intelligence instances, particularly to internal IP ranges, cloud metadata service endpoints (169.254.169.254), or unexpected external hosts. Audit application logs for error messages or debug output indicating failed or successful HTTP requests to unexpected destinations. Review network traffic from the system for patterns consistent with network reconnaissance (rapid connections to multiple hosts on internal subnets). Correlate authentication logs with outbound connection timing to identify potentially compromised accounts. Implement DLP or egress filtering rules that flag requests to sensitive internal services originating from analytics systems.

Why prioritize this

Although the CVSS score is moderate (4.3), prioritize this vulnerability if your environment has sensitive internal services, cloud instances with metadata endpoints, or systems where watsonx.data intelligence has broad network access. The requirement for authentication reduces immediate risk, but compromised user accounts or insider threats make exploitation plausible. Organizations relying on network segmentation to isolate analytics platforms should treat this as higher priority, as SSRF undermines that control.

Risk score, explained

The CVSS 3.1 score of 4.3 reflects: network-accessible attack vector (AV:N), low attack complexity (AC:L), and low privilege requirements (PR:L—authenticated user). Confidentiality impact is low (C:L) because the attacker can read responses to probed requests, gaining information about internal services. Integrity and availability are not directly impacted (I:N/A:N). The scope is unchanged (S:U), meaning impact is confined to the vulnerable component. The moderate score appropriately reflects that while the flaw is real and can facilitate reconnaissance, it does not enable direct data theft, denial of service, or system compromise without chaining to other vulnerabilities.

Frequently asked questions