CVE-2026-0532: Google Gemini Connector SSRF Vulnerability

CVE-2026-0532 Overview

CVE-2026-0532 is a high-severity vulnerability affecting Kibana's Google Gemini connector configuration functionality. The vulnerability combines External Control of File Name or Path (CWE-73) with Server-Side Request Forgery (CWE-918), allowing an attacker to achieve arbitrary file disclosure through a specially crafted credentials JSON payload. This requires authenticated access with privileges sufficient to create or modify connectors (Alerts & Connectors: All). When exploited, the server processes configuration data without proper validation, enabling arbitrary network requests and file reads from the target system.

Critical Impact
Authenticated attackers with connector management privileges can leverage this SSRF and path traversal combination to read arbitrary files from the server and make unauthorized network requests, potentially exposing sensitive configuration data, credentials, and internal network resources.

Affected Products

Kibana versions prior to 8.19.10
Kibana versions prior to 9.1.10
Kibana versions prior to 9.2.4

Discovery Timeline

January 14, 2026 - CVE-2026-0532 published to NVD
January 14, 2026 - Last updated in NVD database

Technical Details for CVE-2026-0532

Vulnerability Analysis

This vulnerability arises from insufficient input validation in Kibana's Google Gemini connector configuration handler. The connector accepts credentials in JSON format, but fails to properly sanitize file path components and URL parameters within the payload. This dual-weakness allows attackers to chain External Control of File Name or Path with Server-Side Request Forgery to achieve arbitrary file disclosure.

The attack requires the attacker to have authenticated access with specific privileges—namely the ability to create or modify connectors through the Alerts & Connectors functionality. While this limits the attack surface to privileged users, in multi-tenant environments or scenarios where lower-privileged administrators have connector management access, this vulnerability poses significant risk.

Root Cause

The root cause stems from inadequate validation of user-controlled input within the Google Gemini connector's credentials JSON payload processing. The server-side code trusts file path and URL components provided in the configuration without sanitizing them against path traversal sequences or validating that network requests are constrained to expected endpoints. This allows malicious actors to craft payloads that reference local file system paths or internal network resources.

Attack Vector

The attack is network-based and requires authentication with Alerts & Connectors: All privileges. An attacker would craft a malicious credentials JSON payload for the Google Gemini connector configuration containing path traversal sequences (such as ../) or internal URLs. When the connector processes this configuration, it follows the attacker-controlled paths and URLs, reading arbitrary files from the file system or making requests to internal services.

The vulnerability mechanism involves crafting a credentials JSON payload with manipulated path or URL parameters. When submitted through the connector configuration interface, the server processes the payload without proper validation. The SSRF component allows the attacker to direct requests to internal services or metadata endpoints, while the path traversal component enables reading sensitive files like /etc/passwd, application configuration files, or credential stores. For detailed technical information, see the Elastic Security Update Announcement.

Detection Methods for CVE-2026-0532

Indicators of Compromise

Unusual file access patterns from the Kibana process, particularly reads of sensitive system files like /etc/passwd, /etc/shadow, or application credential files
Outbound network connections from Kibana to internal metadata services (e.g., cloud provider metadata endpoints at 169.254.169.254)
Connector configuration changes containing path traversal sequences (../, ..\\) or internal IP addresses
Unexpected modifications to Google Gemini connector credentials by users with Alerts & Connectors privileges

Detection Strategies

Monitor Kibana audit logs for connector creation and modification events, filtering for Google Gemini connector types
Implement network monitoring to detect Kibana making requests to internal IP ranges or cloud metadata endpoints
Deploy file integrity monitoring on the Kibana server to alert on unexpected file read operations
Review connector configurations for suspicious patterns including path traversal sequences and internal URLs

Monitoring Recommendations

Enable verbose logging for Kibana connector operations and forward logs to SIEM for correlation
Configure alerts for any Kibana process attempting to access files outside its designated data directories
Monitor for unusual authentication patterns to accounts with Alerts & Connectors: All privileges
Implement egress filtering and monitor for Kibana connections to unexpected internal or external endpoints

How to Mitigate CVE-2026-0532

Immediate Actions Required

Upgrade Kibana to version 8.19.10, 9.1.10, or 9.2.4 or later immediately
Audit all existing Google Gemini connector configurations for suspicious payload content
Review access controls for users with Alerts & Connectors: All privileges and restrict to essential personnel only
Implement network segmentation to limit Kibana's ability to reach sensitive internal services

Patch Information

Elastic has released security updates addressing this vulnerability. Organizations should upgrade to the following patched versions:

Kibana 8.19.10 for the 8.x branch
Kibana 9.1.10 for the 9.1.x branch
Kibana 9.2.4 for the 9.2.x branch

For complete details, refer to the Elastic Security Update Announcement (ESA-2026-05).

Workarounds

Temporarily disable the Google Gemini connector functionality if not required for business operations
Restrict the Alerts & Connectors: All privilege to only trusted administrators who require connector management
Implement web application firewall (WAF) rules to filter requests containing path traversal patterns targeting connector endpoints
Apply network-level controls to prevent Kibana from making outbound requests to sensitive internal resources or metadata endpoints

bash

# Example: Restrict Kibana network access using iptables
# Block access to cloud metadata endpoint
iptables -A OUTPUT -m owner --uid-owner kibana -d 169.254.169.254 -j DROP

# Block access to internal network ranges (adjust as needed)
iptables -A OUTPUT -m owner --uid-owner kibana -d 10.0.0.0/8 -j DROP
iptables -A OUTPUT -m owner --uid-owner kibana -d 172.16.0.0/12 -j DROP
iptables -A OUTPUT -m owner --uid-owner kibana -d 192.168.0.0/16 -j DROP