IBM Langflow Desktop CVE-2026-3340

The vulnerability is exploitable against unauthenticated network access to the Langflow Desktop HTTP endpoint without any required user interaction or special configuration - the SSRF flaw exists in default installations (AV:N/AC:L/PR:N/UI:N). However, practical exploitation is constrained by: (1) the product being Desktop-grade software, not typically exposed to untrusted networks; (2) outbound network policies in enterprise environments that restrict traffic to whitelisted destinations; (3) internal firewalls blocking access to sensitive services (Kubernetes API, cloud metadata services, internal databases) that would be the primary targets of SSRF probing. Exploitation is most feasible in scenarios where Langflow Desktop is exposed via port forwarding, runs in permissive development environments, or is compromised via workstation compromise or insider threat.

CVSS 6.5 with network-accessible, unauthenticated exploitation (AV:N/PR:N) represents moderate-to-high real-world risk for organizations running Langflow Desktop as a service or in shared environments. However, critical limiting factors reduce immediate threat: (1) the product is explicitly 'Desktop' - typically runs on user workstations rather than cloud or production servers, constraining network access to internal corporate networks; (2) CVSS vector shows S:U (scope unchanged), meaning the attacker cannot escalate impact beyond the vulnerable component itself, ruling out lateral movement into other systems via the SSRF alone; (3) no EPSS score or KEV status provided - absence of active exploitation data suggests this is not yet widely weaponized despite public disclosure. The moderate C:L/I:L impact (not full confidentiality/integrity loss) reflects that SSRF enables read access to internal services and triggering of state-changing requests, but not direct data theft or code execution. Organizations using Langflow Desktop in air-gapped or restrictive egress-filtered networks face lower risk; those integrating it with APIs, internal dashboards, or metadata services (AWS IAM, Kubernetes API, internal Jenkins) face elevated risk. Real priority depends on deployment context and network exposure rather than raw CVSS.

An attacker with network access to a Langflow Desktop instance (e.g., exposed via port forwarding, internal network, or compromised user workstation) sends a crafted HTTP request containing a malicious URL parameter to the vulnerable SSRF endpoint. The application processes the URL without validation and initiates an HTTP request from the server, allowing the attacker to: (1) discover internal services by probing common private IP ranges and ports; (2) trigger state-changing requests to internal APIs (disable users, modify configurations, trigger deployments); (3) extract error messages or response content from internal services to identify running software versions or gain information for secondary attacks. No user interaction or authentication is required due to the AV:N/PR:N vector.