CISA adding the Langflow and Trend Micro Apex One vulnerabilities to its Known Exploited Vulnerabilities catalog is a clear signal that these are not theoretical risks anymore. KEV listing means there is evidence of active exploitation, so organizations should treat this as an urgent remediation item rather than another CVE entry waiting patiently in a patch queue nobody wants to look at.

The Langflow vulnerability, CVE-2025-34291, is especially concerning because it can allow arbitrary code execution and full system compromise. The report notes that the issue stems from a combination of permissive CORS, lack of CSRF protection, and an endpoint that allows code execution by design. Successful exploitation can also expose sensitive access tokens and API keys stored in the Langflow workspace, creating a cascading risk across connected SaaS and cloud environments. 

This matters because AI and automation platforms are increasingly connected to business data, cloud accounts, APIs, and internal workflows. If such a platform is compromised, attackers may not just control the application, but also gain access to downstream services through stored tokens. Because apparently the new enterprise attack surface is now “every experimental AI tool someone deployed and forgot to secure.” 

The Trend Micro Apex One issue, CVE-2026-34926, is also serious because it affects on-premise deployments and can allow a pre-authenticated local attacker with server access and admin credentials to modify a key table and inject malicious code for deployment to agents. Trend Micro has observed at least one exploitation attempt in the wild. 
Organizations should immediately identify exposed Langflow and Apex One deployments, apply vendor fixes, rotate potentially exposed API keys and tokens, review administrative access, and inspect logs for signs of exploitation. For Langflow in particular, teams should avoid exposing instances directly to the internet, enforce strong authentication, restrict cross-origin access, and review what secrets are stored inside workspaces.

The broader lesson is simple: security teams need visibility into both traditional security platforms and newer AI/automation tools. Attackers do not care whether a vulnerable system is a production application, a security console, or an AI workflow experiment. If it has access, credentials, and network reach, it is useful to them.

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) on Thursday added two security flaws impacting Langflow and Trend Micro Apex One to its Known Exploited Vulnerabilities (KEV) catalog, citing evidence of active exploitation. The vulnerabilities in question are listed below - CVE-2025-34291 (CVSS score: 9.4) - An origin validation error vulnerability in Langflow that could

Source: CISA Adds Exploited Langflow and Trend Micro Apex One Vulnerabilities to KEV via The Hacker News — published 22 May 2026.