IBM Langflow OSS CVE-2026-90991

Exploitation requires a valid authenticated account on the Langflow OSS instance (CVSS PR:L), so the service must be reachable by the attacker over the network and the attacker must possess or obtain working credentials - this authentication requirement is the primary limiting factor, meaning fully unauthenticated attackers cannot trigger the flaw. Affected deployments are IBM Langflow OSS versions 1.0.0 through 1.12.4. The specific feature, endpoint, or request pattern that permits allocation of resources without limits or throttling is not described in the available sources, so beyond the need for authentication and network reachability no further concrete precondition can be confirmed; impact is limited to availability (no data disclosure or integrity loss per C:N/I:N).

This is a moderate-severity availability issue, not an urgent priority. The CVSS vector (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H) indicates a network-reachable flaw with low attack complexity that nonetheless requires authentication (PR:L), with impact confined to availability (C:N/I:N) - consistent with the CWE-770 uncontrolled resource allocation and the DoS description. No EPSS score is provided, the CVE is not in CISA KEV, and no public exploit or PoC reference is supplied, so there is no public exploit identified at time of analysis and no evidence of active exploitation. The vuln.today decision label is 'this_month', signalling routine prioritization rather than emergency action. The signals are internally consistent: a real but contained risk requiring a valid (low-privilege) account to trigger resource exhaustion against an available service. The main caveat is that the exact triggering mechanism (which endpoint or feature allows unthrottled allocation) is not described in the available sources, so the practical ease of exploitation by an authenticated user cannot be fully confirmed from the data provided.

An attacker who already holds a valid low-privilege account on an exposed Langflow OSS instance (1.0.0-1.12.4) sends requests that trigger resource allocation without limits or throttling, exhausting memory or compute until the service degrades or becomes unavailable to legitimate users. The network attack vector and low complexity make repeated abuse straightforward once authenticated, though no public exploit code is identified at time of analysis.