CVE-2026-103291 in Ghost
Summary
by MITRE • 10/01/2026
Ghost versions from 3.20.2 before 6.51.0 contain a server-side request forgery vulnerability in image dimension refetching that allows authenticated staff users to trigger outbound HTTP requests to arbitrary URLs. Attackers can point image cards at attacker-controlled hosts or internal network endpoints to access metadata services and internal resources unavailable from the public internet.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 10/01/2026
The Ghost content management platform, specifically versions ranging from 3.20.2 up to but not including 6.51.0, contains a critical server-side request forgery vulnerability located within its image dimension refetching mechanism. This flaw arises because the application processes user-supplied URLs for images without adequately validating or restricting the destination hosts before initiating network requests on behalf of the authenticated user. When an administrator or staff member interacts with the platform to update or refresh image metadata, the backend service constructs and executes HTTP GET requests to fetch dimensions such as width and height from the provided URL source. Because this process is triggered by a privileged action but lacks strict allow-listing for internal IP ranges or cloud provider metadata endpoints, it creates an avenue for abuse that bypasses standard network perimeter defenses designed to protect sensitive infrastructure from external access.
This vulnerability allows authenticated staff users with sufficient privileges to manipulate the target of these outbound requests. An attacker who has compromised a staff account can craft malicious image URLs pointing not only to externally controlled domains but also directly to internal network endpoints, such as local loopback addresses or private subnets that are typically isolated from public internet access. The most significant operational impact involves accessing cloud provider metadata services, which often contain sensitive authentication credentials, instance identity documents, and configuration details necessary for further exploitation of the underlying infrastructure. By directing these requests toward internal resources, an attacker can potentially exfiltrate data from databases, API gateways, or other microservices that are not exposed to the public internet but remain reachable within the application's network environment.
From a classification perspective, this vulnerability aligns with CWE-918 Server-Side Request Forgery (SSRF), where the web server makes requests on behalf of the user without proper validation of the destination address. The exploitation technique maps directly to MITRE ATT&CK techniques related to Discovery and Credential Access, specifically leveraging SSRF to probe internal network topology or retrieve instance metadata for lateral movement and privilege escalation. This capability effectively neutralizes network segmentation strategies that rely on keeping certain services hidden from direct external access, as the application itself acts as a proxy into those protected zones.
To mitigate this risk, organizations running affected versions of Ghost must upgrade immediately to version 6.51.0 or later, where the image dimension refetching logic has been hardened against such abuse. In addition to upgrading, administrators should implement strict egress filtering rules at the network level to restrict outbound HTTP and HTTPS traffic from application servers to only known legitimate destinations. This includes blocking access to common cloud metadata endpoints like 169.254.169.254 for AWS or equivalent addresses for other providers. Furthermore, implementing a web application firewall with SSRF detection capabilities can provide an additional layer of defense by inspecting outbound request headers and destination IPs in real-time, ensuring that even if the application logic is flawed, malicious requests are intercepted before they reach internal resources.