As reported by CISA, three vulnerabilities have been added to the Known Exploited Vulnerabilities (KEV) Catalog with confirmed evidence of active exploitation in the wild. This latest batch is particularly notable not just for its technical breadth, but for the strategic value each affected platform represents to threat actors — from AI development infrastructure to managed service provider tooling to foundational Java web serving.

Security Impact: As reported by CISA, three vulnerabilities have been added to the Known Exploited Vulnerabilities (KEV) Catalog with confirmed evidence of active exploitation in the wild.

Vulnerability Breakdown

CVE-2026-9198 — IBM Langflow Code Injection
Vulnerability Breakdown
Type: Code Injection
Affected Product: IBM Langflow (AI workflow orchestration platform)
Risk: Remote code execution; full system compromise post-exploitation
Active Exploitation: Confirmed by CISA
Patch Status: Check IBM's security advisory portal for the latest patch guidance
CVE-2026-18556 — N-able N-central Authentication Bypass
  • Type: Authentication Bypass via Alternate Path or Channel
  • Affected Product: N-able N-central (MSP remote monitoring and management platform)
  • Risk: Unauthenticated access to management console; lateral movement into managed client environments
  • Active Exploitation: Confirmed by CISA
  • Patch Status: Consult N-able's security bulletin for remediation steps
CVE-2026-34486 — Apache Tomcat Missing Encryption of Sensitive Data
  • Type: Missing Encryption of Sensitive Data (CWE-311)
  • Affected Product: Apache Tomcat (widely deployed Java servlet container)
  • Risk: Interception of sensitive data in transit; credential theft; session hijacking
  • Active Exploitation: Confirmed by CISA
  • Patch Status: Refer to the Apache Tomcat security pages for patched versions

Why This Cluster of Vulnerabilities Matters

What makes this KEV addition especially significant is the diversity of the affected attack surface. These three vulnerabilities don't just hit one sector — they collectively expose AI/ML development pipelines, the managed services supply chain, and core web application infrastructure simultaneously.

The IBM Langflow code injection vulnerability is a particular area of concern from Shield53's perspective. Langflow is an AI workflow orchestration tool increasingly deployed in enterprise environments to build LLM-based applications. A code injection flaw in this kind of platform isn't just a server compromise — it's a potential backdoor into AI pipeline logic, training data handling, and API key stores. Organizations building internal AI tooling should treat this with the urgency of a critical infrastructure compromise, not a routine patch cycle.

The N-able N-central authentication bypass carries enormous supply chain risk. MSPs use N-central to manage hundreds or thousands of customer endpoints. A threat actor who can bypass authentication on an MSP's N-central instance doesn't just own that MSP — they potentially own every client managed through that platform. This mirrors the attack pattern seen in historic MSP-targeting campaigns and should immediately trigger a review of any N-central deployment exposed to the internet.

Shield53 Analyst View: MSP-targeting vulnerabilities consistently yield the highest attacker ROI. One compromised management console can cascade into dozens of downstream victims. If you are an MSP or a client of an MSP using N-able N-central, assume risk until patched and audited.

The Apache Tomcat missing encryption vulnerability may appear less dramatic than code execution, but in practice it enables credential interception and session compromise at scale. Tomcat underpins a vast number of enterprise Java applications — including ERP frontends, internal portals, and legacy business applications — making this a wide-blast-radius issue that defenders often underestimate.

Who Is Most at Risk

  • AI/ML Engineering Teams: Any organization running IBM Langflow in development or production environments, particularly those with internet-facing instances
  • Managed Service Providers and Their Clients: N-able N-central users, especially those with externally accessible management interfaces
  • Enterprises Running Java Web Applications: Apache Tomcat deployments across healthcare, finance, government, and manufacturing sectors
  • Federal Civilian Agencies: Bound by BOD 26-04 remediation timelines with mandatory patch windows

Shield53 Recommendations — Immediate Actions

For CVE-2026-9198 (IBM Langflow)

  • Immediately identify all Langflow deployments across your environment, including developer sandboxes and CI/CD pipelines
  • Restrict internet exposure of Langflow interfaces; place behind VPN or Zero Trust access controls pending patch
  • Audit API keys, credentials, and secrets accessible within Langflow workflows
  • Apply IBM's patch as soon as it is available; monitor IBM's PSIRT portal for updates
  • Review logs for unusual code execution activity or unexpected outbound connections from Langflow hosts

For CVE-2026-18556 (N-able N-central)

  • Immediately restrict internet-facing access to N-central management consoles
  • Review N-central audit logs for unauthorized access attempts or configuration changes
  • Apply N-able's patch immediately upon availability; treat this as a P1 incident given supply chain blast radius
  • Notify downstream managed clients that a security review is underway and watch for anomalous activity across managed endpoints
  • Enable MFA on all N-central accounts regardless of patch status

For CVE-2026-34486 (Apache Tomcat)

  • Inventory all Apache Tomcat instances across on-premises and cloud environments
  • Upgrade to the patched Tomcat version immediately; consult the Apache Tomcat security advisories for the correct target version
  • Enforce TLS 1.2+ across all Tomcat connectors and disable legacy unencrypted channels
  • Review connector configuration files (server.xml) for missing or misconfigured SSL/TLS settings
  • Inspect network traffic for signs of credential interception or anomalous authentication patterns

Broader Vulnerability Management Guidance

  • All organizations — not just federal agencies — should align their patch prioritization with the CISA KEV Catalog. KEV entries represent confirmed real-world exploitation, making them the highest-priority items in any vulnerability management program.
  • Implement continuous asset discovery to ensure shadow IT and unmanaged instances of affected products are identified quickly.
  • For organizations that have not yet adopted a KEV-aligned patching SLA, Shield53 recommends a 72-hour critical response window for any KEV addition affecting internet-exposed assets.