CVE-2026-9198

IBM Langflow OSS 1.0.0 through 1.10.0 contains an unauthenticated code-injection flaw (CWE-94) that can be chained into full remote code execution. An attacker with network access first calls /api/v1/auto_login, which mints a SUPERUSER token to any caller without authentication, then calls /api/v1/validate/code, which executes attacker-supplied code via exec() on the Langflow host. Successful exploitation yields full control of the affected deployment, with high impact on confidentiality, integrity, and availability (CVSS 3.1: 9.8, network-exploitable with no privileges or user interaction required). Anyone running IBM Langflow OSS 1.0.0-1.10.0 is affected, especially default deployments reachable from untrusted networks. CISA added the flaw to the Known Exploited Vulnerabilities catalog on 2026-08-04, confirming active exploitation; EPSS estimates a 60.6% probability of exploitation within 30 days (99th percentile), and ransomware use is unknown.

What to do: Upgrade IBM Langflow OSS to a release newer than 1.10.0 per IBM's ([email protected]) security advisory, as required under CISA's BOD 26-04 guidance for KEV entries; if mitigations are unavailable, discontinue use of the product per BOD 26-04. Until patched, restrict network access to the /api/v1/auto_login and /api/v1/validate/code endpoints (e.g., firewall rules or reverse-proxy controls) and avoid exposing the instance directly to the internet. Because exploitation is confirmed, review access logs for requests to these endpoints and signs of unauthenticated SUPERUSER token issuance or unexpected code execution.