In the four weeks between 19 March and 20 April 2026, CISA's Known Exploited Vulnerabilities catalog received a run of entries with something in common: they hit the systems that sit at the edge of a network or manage it. Cisco Secure Firewall Management Center, F5 BIG-IP APM, Citrix NetScaler configured as a SAML identity provider, Fortinet FortiClient EMS twice, Ivanti Endpoint Manager Mobile and three flaws in Cisco Catalyst SD-WAN Manager. Each one carried a three-day remediation deadline for US federal agencies. Looking back from November, spring 2026 was a preview of the whole year: management planes and identity features on internet-facing devices were exploited again and again. This article explains what the KEV catalog is, summarises the spring cases, and turns them into a routine for answering the only question that matters in the first hours: are we exposed? For the broader process, see our guide to the vulnerability management lifecycle.
Key takeaways
- A CVE in the KEV catalog has evidence of exploitation in the wild, which makes it more urgent than a "critical" score alone
- Spring 2026's KEV entries concentrated on management servers and identity features of edge devices
- FortiClient EMS showed why "we upgraded" is not the end: the version that fixed one exploited flaw was hit by the next
- The first question is not "is it critical?" but "do we expose this product, in which version, and where?"
- Patching closes the hole; it does not remove an attacker who came in before the patch
What the KEV catalog is, and why "exploited" outranks "critical"
The Known Exploited Vulnerabilities catalog is maintained by CISA, the US Cybersecurity and Infrastructure Security Agency. A CVE enters it when there is reliable evidence that attackers are using it, a clear remediation exists and the CVE has been assigned. US federal agencies must remediate within the due date set for each entry; for everyone else, including European companies, the catalog is a free, machine-readable list of what attackers are actually using.
The difference from a severity score is the point. CVSS measures how bad a vulnerability would be if exploited. It says nothing about whether anyone is exploiting it. Every year tens of thousands of CVEs are published, and a large share are rated high or critical; only a small fraction ever appear in KEV. A medium-severity flaw in KEV on a device you expose to the internet deserves more attention than a critical one on an internal system no attacker can reach. KEV also flags entries known to be used in ransomware campaigns.
The spring 2026 cases in one table
The table below lists the edge and management-plane entries added to KEV between 19 March and 20 April 2026, with the vendor advisory linked in each KEV record.
| Product | CVE | Flaw | Added to KEV | Notes |
|---|---|---|---|---|
| Cisco Secure Firewall Management Center | CVE-2026-20131 | Deserialization in the web management interface: unauthenticated remote code execution as root | 19 March | Advisory published 4 March, CVSS 10.0, no workarounds; ransomware use "Known" |
| F5 BIG-IP APM | CVE-2025-53521 | Stack-based buffer overflow leading to remote code execution | 27 March | Access policy module, often the remote access entry point |
| Citrix NetScaler ADC and Gateway | CVE-2026-3055 | Memory overread when configured as a SAML identity provider | 30 March | Bulletin CTX696300 published 23 March; default configurations not affected |
| Fortinet FortiClient EMS | CVE-2026-35616 | Improper access control: unauthenticated code or command execution | 6 April | Advisory FG-IR-26-099 of 4 April, CVSS 9.8; versions 7.4.5 and 7.4.6; hotfix, then 7.4.7 |
| Ivanti Endpoint Manager Mobile | CVE-2026-1340 | Code injection: unauthenticated remote code execution | 8 April | Mobile device management server |
| Fortinet FortiClient EMS | CVE-2026-21643 | SQL injection: unauthenticated code or command execution | 13 April | Advisory FG-IR-25-1142 of 6 February; version 7.4.4, fixed in 7.4.5 |
| Cisco Catalyst SD-WAN Manager | CVE-2026-20122, -20128, -20133 | File overwrite through the API, recoverable stored passwords, information exposure | 20 April | Three entries the same day on the SD-WAN management plane |
Sources: CISA KEV catalog; Cisco advisory for CVE-2026-20131; FG-IR-26-099; FG-IR-25-1142; Citrix bulletin CTX696300.
FortiClient EMS: when the fix leads to the next flaw
The two FortiClient EMS entries are worth reading together. Fortinet's February advisory for CVE-2026-21643 told customers on 7.4.4 to upgrade to 7.4.5 or above; Fortinet's advisory reports that flaw as exploited in the wild, and CISA added it to KEV on 13 April. Meanwhile, on 4 April, Fortinet had published CVE-2026-35616, which affects exactly 7.4.5 and 7.4.6, also exploited, with a hotfix to apply until 7.4.7 was available. An organisation that did the right thing in February found itself in range of the next exploited flaw two months later. Upgrading is a step, not a state: the product stays on your watch list for as long as it faces the internet.
The pattern: management planes and identity features
None of these devices are exotic. They are the firewall manager, the SD-WAN controller, the server that pushes policies to every laptop, the device management server for every phone, the gateway that authenticates remote users. Three things make them attractive:
- They hold the keys to everything else. A firewall management center configures every firewall; an endpoint management server can push software to every endpoint; an SD-WAN manager controls the whole WAN. Compromising the manager is worth more than compromising any one managed device.
- They are often reachable from the internet when they need not be. Endpoint and mobile management servers are exposed so that remote devices can check in; management interfaces end up exposed for convenience, for a supplier, or by mistake.
- They carry identity functions. A NetScaler acting as a SAML identity provider or a BIG-IP access policy module sits in the authentication path. A flaw there can leak session data or bypass access controls for every application behind it.
Note also what the Citrix case teaches: CVE-2026-3055 affected only appliances configured as a SAML identity provider. "We have NetScaler" is not enough to decide whether you are exposed; you need to know which features are switched on.
The first question: do we expose it, which version, where?
When a KEV entry for an edge product appears, the useful answer in the first hours is a list: every instance of that product reachable from the internet, its version, the feature that matters, and an owner. Most organisations cannot produce it quickly, for three reasons:
- The asset register lists what IT installed, not what is reachable. Devices installed by a supplier, a subsidiary or a project often are not there.
- The version recorded in the register is the one at installation time, not the one running today.
- Management interfaces live on unusual ports or forgotten hostnames, and nobody looks for them.
The fastest reliable answer comes from the outside: discover every host under your domains and IP ranges, identify the products by banners, login pages and fingerprints, and match them against the KEV entry. That is the same view the attacker has, and attackers run it against the whole internet within hours of a disclosure.
Patch, then check for compromise
When the flaw was exploited before you patched, patching closes the door but does not tell you whether someone already walked through it. CISA's note on the FortiClient EMS entry is explicit: check for signs of potential compromise on all internet-accessible affected products. In practice:
- Apply the fix or the vendor mitigation and confirm the running version afterwards.
- Review the device for changes you did not make: new administrator accounts, configuration changes, new scheduled tasks or scripts, unexpected files.
- Review logs from before the patch for access to the management interface from unknown addresses, and keep them: some devices retain logs only for a short time.
- Rotate secrets stored on or passing through the device: administrator passwords, API keys, service accounts used to talk to directories, certificates. A management server that was compromised has seen them.
- Look downstream. If the device manages others, check what it pushed to them in the exposure window.
- Escalate early to incident response when you find any sign of compromise. Rebuilding from a known-good image is often safer than cleaning.
A 24 to 72 hour routine for KEV
The three-day federal deadline in every spring entry is a good benchmark for any organisation. A routine that meets it looks like this:
- Hour 0 to 4: know. Subscribe to the KEV feed (it is published as JSON and CSV) and to the security advisories of your edge vendors. Someone is responsible for reading them every working day.
- Hour 4 to 24: match and triage. Match the new entry against an inventory measured from outside. For every exposed instance, decide: patch now, mitigate now, or take it off the internet now.
- Hour 24 to 72: fix and verify. Apply the fix through an emergency change process agreed in advance. Then verify from the outside that the vulnerability is gone, not just that the patch job reported success.
- In parallel: check for compromise as described above, starting with the most exposed instances.
- Afterwards: reduce. Ask whether the management interface needs to be on the internet at all. Allow-lists, a dedicated management network and VPN-only access remove whole classes of future emergencies; our article on reducing the attack surface covers the options.
- Document. For NIS2 entities, Article 21(2)(e) covers vulnerability handling. A record showing when you learned of a KEV entry, when you were exposed and when you fixed it is exactly the evidence an auditor will ask for.
How Orizon helps
RECON External answers the first question from the outside, starting from your domain name. It finds every subdomain and host, which ports answer and what each service says about itself, and recognises admin and login panels by their favicon even on unusual ports. Detected products are matched to known CVEs and weighed by CISA KEV, Exploit-DB, Metasploit and EPSS. Then each finding is tested against the target from the scan's own isolated server and marked Proven, To verify or Closed, with the evidence of every test; actively exploited CVEs are decided on the evidence like every other finding. RECON confirms that a vulnerability exists; it does not exploit it. Scans can be scheduled daily, so a new KEV entry is matched against your current surface, not last quarter's.
If you need an assessment with a report for management or an auditor, our vulnerability assessment service is built on the same verification. And when you need impact proven on the external perimeter you authorise, Fireline takes it from there. To understand the difference, read vulnerability assessment vs penetration testing.