Ghost CMS CVE-2026-53944

Exploitation requires that the attacker can supply a URL to a Ghost feature that performs server-side outbound HTTP requests (such as bookmark/embed URL previews, newsletter fetches, or webhook configurations). Ghost versions 6.0.9 through 6.21.0 must be running. No authentication is required per the CVSS PR:N rating. The attacker must know or discover a Ghost endpoint that accepts attacker-controlled URLs for server-side fetching. The IPv6 literal must use the ::ffff:<private-IPv4> mapped format to bypass the blocklist. Exploitation is bounded by what internal services are network-reachable from the Ghost host - deployments with strict outbound firewall rules blocking RFC 1918 destinations are partially protected even without the patch. Non-cloud deployments with no sensitive internal services face reduced real-world impact despite the bypass being equally achievable.

The CVSS 3.1 base score of 5.8 (Medium) encodes AV:N/AC:L/PR:N/UI:N/S:C/C:N/I:L/A:N. The scope change (S:C) is the most operationally significant metric: successful exploitation crosses the trust boundary from the Ghost application into adjacent internal services. The vendor-assigned C:N is conservative - SSRF to internal metadata endpoints (e.g., cloud IMDSv1 at 169.254.169.254) or internal APIs commonly yields credential or configuration disclosure, suggesting real-world confidentiality impact may exceed the scored value. EPSS data was not provided, preventing probabilistic quantification of exploitation likelihood. The vulnerability is not listed in CISA KEV, and no public exploit code has been identified. Despite the medium CVSS score, deployments on cloud infrastructure with accessible metadata services face elevated risk because the bypass technique (IPv6-mapped literals) is trivially constructed and the attack requires no authentication per PR:N. Deployments behind strict egress firewalls are meaningfully protected.

An unauthenticated attacker identifies a Ghost feature that triggers server-side outbound HTTP requests - such as a URL preview, bookmark card, or webhook - and submits a URL using an IPv6-mapped private IPv4 literal (e.g., http://[::ffff:169.254.169.254]/latest/meta-data/iam/security-credentials/). Ghost's IP filter passes the address as non-private, the Node.js runtime connects to the cloud metadata service, and the attacker receives IAM credential material in the response. No public exploit code has been confirmed at time of analysis, but the technique is well-understood, requires no tooling beyond a standard HTTP client, and is trivially constructed.