IBM Langflow OSS EUVD-2026-113135

Exploitation requires a valid low-privileged authenticated session against a Langflow OSS instance running an affected version (1.0.0 through 1.12.4), reachable over the network; no user interaction is needed (AV:N/AC:L/PR:L/UI:N). The specific product feature or code-generation pathway that is improperly controlled is not described in the available sources beyond CWE-94 (improper control of code generation). The primary limiting factor is the authentication requirement - an attacker must first hold or obtain account access, so this is not a pre-authentication or anonymous-internet exploit. Applying the vendor patch is the stated remediation.

This is a genuinely serious vulnerability that warrants prompt but not emergency action, consistent with the input's 'this_week' decision label. The CVSS vector (AV:N/AC:L/PR:L/UI:N) describes a network-reachable, low-complexity flaw requiring only low-privileged authentication and no user interaction, yielding total compromise (C:H/I:H/A:H) - arbitrary code execution via improper control of code generation (CWE-94). The SSVC framework tempers raw severity: Exploitation is 'none' (no observed or public exploitation), Automatable is 'no' (not amenable to mass, automated weaponization), while Technical Impact is 'total' - so the harm ceiling is maximal but the likelihood signals are currently low. There is no CISA KEV listing and no EPSS score was supplied, so exploitation probability cannot be quantified from the input. No POC or exploit reference appears in the data. Net: a high-impact, low-barrier authenticated RCE in an AI application-building platform that should be patched this week, but current signals point to latent rather than active risk. The authentication requirement (PR:L) is the main brake - this is a post-authentication flaw, not a pre-auth internet-wide crisis.

An attacker with low-privileged but valid access to a Langflow OSS instance (1.0.0-1.12.4) abuses the improper control of code generation to inject and execute arbitrary code on the server. Given AC:L and no required user interaction, a logged-in user - or an attacker who has obtained any low-tier credential - can trigger it reliably over the network. No public exploit code was identified in the provided sources at time of analysis.