FreeRDP EUVD-2026-101718

Exploitation requires the FreeRDP proxy to be deployed with FixedTarget=false AND to forward connections through an upstream HTTP proxy; only in this mode is the attacker-influenced routing token target (Cookie: msts=) written into the outgoing CONNECT request line and Host header. The limiting factors are significant: the default/common FixedTarget=true mode and deployments without an upstream HTTP proxy are not affected, and the attacker must be able to control the routing token target. Impact is confined to HTTP request/header framing manipulation at the upstream proxy (integrity only) - no code execution, data disclosure, or availability loss is implied by the data.

This is a genuinely medium-tier issue, not an inflated high-CVSS item - the provided CVSS 4.0 vector (AV:N/AC:H/AT:P/PR:N/UI:N) scores 6.3 and the impact is limited to low integrity on both the vulnerable system (VI:L) and a subsequent system (SI:L), with no confidentiality or availability effect. The attack is network-reachable and unauthenticated (PR:N), which raises concern, but this is offset by AC:H and AT:P, meaning exploitation depends on a specific non-default deployment (FreeRDP running as a proxy with FixedTarget=false in front of an upstream HTTP proxy) and on the attacker being able to influence the routing token. No EPSS score, KEV listing, or POC was supplied in the intelligence; per precision rules this is 'no public exploit identified at time of analysis.' There is no SSVC decision data provided. The signals are internally consistent - network vector matches the described remote RDP client connection, and the low/low integrity impact matches an HTTP request/header-smuggling primitive rather than code execution. The main caveat: the real-world exposure is gated entirely by the proxy configuration, so the network vector overstates risk for the overwhelming majority of FreeRDP deployments that do not run the proxy in this mode.

An attacker operates or targets a FreeRDP proxy configured with FixedTarget=false that forwards through an upstream HTTP proxy. As a remote unauthenticated RDP client, the attacker crafts a routing token (Cookie: msts=) whose target portion contains a lone LF, which http_proxy_connect copies unsanitized into the CONNECT request line and Host header, allowing the attacker to split/smuggle HTTP framing at the upstream proxy. No public exploit code was identified in the provided data.