Progress MarkLogic EUVD-2026-97513

Exploitation requires (1) MarkLogic Server running a vulnerable build - 12.x before 12.1.0 or 11.x before 11.3.7 - AND (2) that the instance is deployed on Microsoft Azure, since the SSRF derives its value from reaching Azure internal cloud services and the Instance Metadata Service for credential/token theft. The attacker must also hold valid authentication to the MarkLogic Server (CVSS PR:L), so this is not an unauthenticated flaw. Limiting factors that reduce real-world risk: non-Azure deployments (on-premises or other clouds) are outside the described scope, a working low-privilege credential is a prerequisite, and impact is confidentiality-focused (credential/token disclosure) rather than direct code execution or availability loss.

This is a genuine mid-to-high priority issue but not a drop-everything emergency. The CVSS 3.1 vector (AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:L/A:N, base 8.5) indicates network-reachable, low-complexity exploitation that requires only low-level authentication (PR:L) and no user interaction, with a scope change (S:C) that lets the attacker pivot from the MarkLogic application context into the underlying Azure cloud infrastructure - consistent with a CWE-918 SSRF reaching the Azure Instance Metadata Service. The high confidentiality impact (C:H) reflects theft of cloud credentials/tokens, which is the real danger: it can enable lateral movement and privilege escalation within the victim's Azure tenant. Mitigating factors: exploitation is gated behind valid authentication (PR:L) and is limited to MarkLogic instances deployed on Azure, which narrows the exposed population. No EPSS score was provided in the input, so real-world exploitation probability cannot be quantified from this data. There is no CISA KEV listing and no public POC indicated, so there is no evidence of active exploitation - the 8.5 score reflects potential impact, not observed attacks. The signals are internally consistent (no CVSS-vs-description conflict); the main gap is the missing EPSS/exploitation-maturity telemetry, so prioritization should lean on the credential-theft impact and the authenticated-but-low-privilege barrier rather than on any confirmed exploitation signal.

An attacker who holds a low-privilege authenticated account on an Azure-hosted MarkLogic Server (11.x before 11.3.7 or 12.x before 12.1.0) submits a request that coerces the server into making an attacker-controlled outbound call to an internal cloud endpoint such as the Azure Instance Metadata Service. The server returns the response, exposing managed-identity tokens or cloud credentials, which the attacker reuses to reach internal cloud services within the tenant. No public exploit was identified at time of analysis, so weaponization would require custom development.