MarkLogic Server CVE-2026-85954

Exploitation requires the ability to supply user-controlled values to certain higher-order XQuery functions that MarkLogic Server evaluates in an elevated security context - i.e., access to an interface that evaluates XQuery (a MarkLogic account with query/evaluation rights, or an application that passes user input into such functions). Only MarkLogic Server 12.x before 12.1.0 and 11.x before 11.3.7 are affected, which limits exposure to unpatched instances. The key limiting factor is that the attacker must already possess some granted permissions and a path to submit XQuery; the vendor vector's PR:N (unauthenticated) claim is inconsistent with the description's wording and should be confirmed before assuming no authentication is needed.

The vendor CVSS 3.1 vector (AV:N/AC:L/PR:N/UI:N/S:C/C:N/I:H/A:N, 8.6 High) frames this as a network-reachable, low-complexity, unauthenticated privilege-escalation flaw with a scope change and high integrity impact (no confidentiality or availability impact). However, these signals partially conflict: the description states the escalation goes 'beyond the attacker's granted permissions,' which implies the attacker already holds some MarkLogic account/permissions and the ability to submit XQuery - that reads more like an authenticated, low-privilege precondition (PR:L) than the PR:N in the vendor vector. This discrepancy should be verified with Progress. No EPSS score, CISA KEV listing, or public POC was provided in the input, so there is no exploitation-probability or active-exploitation signal to corroborate urgency. CWE-269 (Improper Privilege Management) and the scope change (S:C) are consistent with a genuine trust-boundary break, so for organizations running affected MarkLogic versions with multiple privilege tiers this is a real, patch-worthy priority - but in the absence of EPSS/KEV/POC data it currently reflects a high-severity design flaw rather than confirmed widespread exploitation. Treat the PR:N (unauthenticated) claim as unconfirmed pending vendor clarification.

A low-privileged MarkLogic user (or an application endpoint that forwards user input into XQuery) supplies crafted values to a higher-order XQuery function that MarkLogic evaluates in an elevated security context. Because the evaluation runs with elevated permissions, the attacker's input executes beyond their granted rights, letting them modify data or perform operations reserved for higher-privilege roles. No public exploit was identified at time of analysis, and the network attack vector with low complexity means no special timing or local access is needed once query submission is possible.