On August 5, 2026, the U.S. Cybersecurity and Infrastructure Security Agency (CISA) added three vulnerabilities to its Known Exploited Vulnerabilities (KEV) catalog. The agency’s decision was based on evidence that attackers are actively exploiting these weaknesses in real-world environments. For many organizations, this is a clear signal that patching cannot wait.
Below is a practical breakdown of the three newly listed CVEs—what they affect, why they matter, and how the timelines shape your risk management.
Why the KEV catalog update matters
The KEV catalog is CISA’s public list of security flaws known to be exploited in the wild. When a vulnerability enters this catalog, it typically means defenders should treat it as a near-term priority because threat actors have already demonstrated an operational use of the bug.
CISA also ties these updates to compliance expectations for Federal Civilian Executive Branch (FCEB) agencies. In this case, agencies have until August 7, 2026 to apply the necessary fixes and reduce exposure.
Langflow CVE-2026-9198: unauthenticated code injection leading to RCE
The first newly added item is CVE-2026-9198 (CVSS 9.8). This issue is described as a code injection vulnerability in Langflow that can allow unauthenticated attackers to achieve full remote code execution on default Langflow deployments.
According to the report, CISA notes that a fix was released in July 2026, with Langflow version 1.10.1 mentioned as the remediation point. The high severity and the fact that exploitation does not require authentication make this one especially urgent.
What’s known about exploitation
CISA did not include detailed information about how this Langflow flaw is being exploited. However, the update points to a broader pattern: security defects in open-source AI application development platforms have repeatedly been leveraged by malicious actors in recent months.
Apache Tomcat CVE-2026-34486: encryption bypass via missing protection
The second newly added vulnerability is CVE-2026-34486 (CVSS 7.5). This one targets Apache Tomcat and is described as a missing encryption of sensitive data weakness.
In practical terms, the flaw allows attackers to bypass EncryptInterceptor, a component used in a cluster to apply pre-shared key encryption to messages exchanged between cluster nodes. If sensitive data is not properly encrypted, defenders should expect increased risk around confidentiality and internal traffic manipulation.
Remediation timeline and versions
CISA states that a fix was issued in April 2026, with updates provided for multiple release lines: 11.0.21, 10.1.54, and 9.0.117.
Attribution and exploitation approach
Unlike the Langflow item, exploitation of CVE-2026-34486 is tied to an AI-enabled autonomous hacking effort. The activity is attributed to a Chinese-speaking threat actor known by aliases including knaithe and KnYuan, reportedly operating from Zhuhai, China.
Reportedly, the actor leveraged DeepSeek through the Hermes Agent framework to act as an offensive operator against internet-exposed devices. The reported workflow included attempts to exploit a different Langflow flaw first—CVE-2026-33017 (CVSS 9.8)—but those initial attempts allegedly failed due to restrictive target configurations.
After that, the AI agent reportedly performed autonomous research to locate other higher-value weaknesses, including additional targets such as flaws in n8n, to find an effective path to compromise.
N-able N-central CVE-2026-18556: authentication bypass and a fresh patch
The third vulnerability in the KEV catalog update is CVE-2026-18556 (CVSS 8.2) affecting N-able N-central. CISA describes this as an authentication bypass issue.
A key detail is that the fix for this problem was not complete at first, which led N-able to release a new patch. That subsequent patched state is tracked as CVE-2026-18577 (CVSS 8.2).
For defenders, this means you should not only check whether a vendor update exists, but also confirm that you are on the latest corrected version associated with the follow-up CVE.
Signals that threat actors are already active
The update indicates that while CVE-2026-18577 was placed in the KEV catalog on Monday, recent development suggests that both N-central vulnerabilities are being exploited by threat actors.
Autonomous plus manual operations: what defenders can take away
Security reporting attributed the campaign’s behavior to a combination of automated and human-driven methods. The threat actor is said to have used both autonomous and manual techniques, including:
- AI-assisted target identification and narrowing to reduce wasted effort and focus on likely-good targets
- Manual operations against other known-vulnerable systems
One quoted assessment noted the actor attempted exploitation across more than 460 targets. It also highlighted that the actor appeared to use DeepSeek to narrow targeting scope, potentially to conserve AI compute resources.
Another notable observation was operational speed: the system reportedly executed hundreds of hours of manual targeting analysis in minutes while managing its compute resources.
What FCEB agencies—and other organizations—should do now
CISA states that FCEB agencies have until August 7, 2026 to apply the required fixes. Even if you are outside the FCEB scope, the KEV catalog update suggests a similar urgency: vulnerabilities listed here are already being weaponized.
To respond effectively, consider the following immediate actions:
- Verify exposure: confirm whether you run Langflow, Apache Tomcat cluster components, and N-able N-central in the affected configurations.
- Patch quickly: align your upgrades with the specific versions cited by CISA (Langflow 1.10.1; Tomcat 11.0.21 / 10.1.54 / 9.0.117; and the N-able fix associated with the follow-up tracking).
- Re-check after changes: ensure the remediation is actually in place and that systems reflect the fixed release state, not just “an update was attempted.”
- Review internet exposure: the reporting focuses heavily on internet-exposed devices, so limit exposure where possible and monitor for suspicious access patterns.
Even though CISA did not provide detailed exploit mechanics for the Langflow issue, the combination of a KEV designation and the reported exploitation of related targets underscores that attackers are actively looking for reachable systems.
Conclusion: treat the KEV catalog as a patch priority list
The latest CISA KEV catalog update spotlights three weaknesses with clear real-world danger: a Langflow unauthenticated code injection to RCE, an Apache Tomcat encryption-bypass path, and an N-central authentication bypass with follow-up patching tracked under another CVE. Because the catalog is specifically built around known exploited issues, organizations should move these items to the front of their vulnerability management queue.
If you haven’t already, start by inventorying affected software and then mapping installed versions to the fixed releases mentioned by CISA. That approach gives you the fastest path to reduce exposure—before attackers widen their reach.
Source: https://thehackernews.com/2026/08/cisa-flags-langflow-rce-tomcat-and-n.html
