Skip to content

Patch what attackers actually use

Machine verified A machine ran this on our install and checked the result against the CLI. A test report is linked. Proven against quoxcli 0.6.0 on 2026-11-10.

What do we patch first?

Thousands of CVEs, one change window. Rank them by real-world exploitation, not by the biggest number.

What you get: the vulnerabilities on your servers in the order that matters. First the ones on the CISA list of flaws known to be exploited in the wild (KEV), then the ones with a high chance of being exploited soon (FIRST EPSS), then the ones that only score high.

quox · talking to Quox

Which vulnerabilities on our servers should we patch first, and why? Tell me plainly which are known to be exploited in the wild and which just have a high score. Don't change anything.

▤ AOCL │ L1 identity complete · allowed · 1ms

▤ AOCL │ L3 policy gate complete · proceed · 3ms

▤ AOCL │ handoff narration kept out of the answer · I'll hand this to ARGUS, our Wazuh specialist, as a read-only request.

◐ DELEGATE │ delegate_to_agent agent_id=argus ✔ DELEGATE │ delegate_to_agent agent_id=argus │ 4 tools used, 4 tool outcomes · 55.7s

◐ ARGUS › WAZUH │ cve priority ✔ ARGUS › WAZUH │ cve priority │ 822ms

◐ ARGUS › WAZUH │ vulnerabilities ✔ ARGUS › WAZUH │ vulnerabilities │ 43ms

◐ ARGUS › WAZUH │ vulnerabilities ✔ ARGUS › WAZUH │ vulnerabilities │ 81ms

◐ ARGUS › WAZUH │ vulnerabilities ✔ ARGUS › WAZUH │ vulnerabilities

⛓ AEE │ envelope opened · ops.tool · delegate_to_agent ok

⛓ AEE │ envelope opened · stream_success

I made no changes and only ran read queries. Seven CVEs on your two monitored servers 
are on the CISA Known Exploited list, and these should be patched first. Your
highest-scoring CVEs (CVSS 10.0) are not on that list and have a low chance of being
exploited, so they come after.

Wazuh holds 54,368 vulnerability records in total. 8,062 are distinct CVEs, and 2,002 of
those were ranked; the ranker marks that as cut off (source: wazuh_cve_priority,
wazuh_vulnerabilities). Both outside data sources answered. The CISA Known Exploited
list (KEV) found 7 matches. The EPSS score, which estimates the chance of exploitation
in the next 30 days, put 10 CVEs above 10% (source: wazuh_cve_priority).


1. Known exploited in the wild (CISA KEV), patch first

The ranks are the ranker's own order. I have not re-sorted them.

Rank               1
CVE                CVE-2026-24061
Host (agent)       hyp-db-01 (002)
Package            inetutils-telnet, inetutils-ping, telnet
CVSS               9.8
EPSS               99.0%
Why it ranks here  Known exploited (added 2026-01-26, CISA deadline 2026-02-16, long
                   passed). Highest EPSS in the fleet.

Rank               2
CVE                CVE-2026-31431
Host (agent)       hyp-db-01 (002), hyp-web-01 (001)
Package            linux-image 6.8.0-88 / 6.8.0-90
CVSS               7.8
EPSS               98.1%
Why it ranks here  Known exploited (deadline 2026-05-15). Very high EPSS and on both
                   hosts.

Rank               3
CVE                CVE-2026-111497
Host (agent)       hyp-db-01 (002)
Package            libbrotli1
CVSS               8.1
EPSS               82.6%
Why it ranks here  Known exploited, added 2026-10-27. The CISA deadline is today,
                   2026-11-10.

Rank               4
CVE                CVE-2026-46300
Host (agent)       both hosts
Package            linux-image 6.8.0-88 / 6.8.0-90
CVSS               7.8
EPSS               16.1%
Why it ranks here  Known exploited (deadline 2026-10-31, passed), on 2 hosts.

Rank               5
CVE                CVE-2025-39964
Host (agent)       hyp-db-01 (002)
Package            linux-image-6.8.0-88
CVSS               5.5
EPSS               1.6%
Why it ranks here  Known exploited even though its score is medium. That is why it ranks
                   above higher-scoring CVEs.

Rank               6
CVE                CVE-2026-53266
Host (agent)       both hosts
Package            linux-image 6.8.0-88 / 6.8.0-90
CVSS               7.0
EPSS               0.9%
Why it ranks here  Known exploited (deadline 2026-09-21), on 2 hosts.

Rank               7
CVE                CVE-2026-53362
Host (agent)       both hosts
Package            linux-image 6.8.0-88 / 6.8.0-90
CVSS               7.8
EPSS               0.7%
Why it ranks here  Known exploited (deadline 2026-08-30), on 2 hosts.

None of the seven is flagged as used in ransomware campaigns (source:
wazuh_cve_priority).

Not on the KEV list but very likely to be exploited, so patch these next:

• Rank 8, CVE-2026-43284, kernel on both hosts: EPSS 94.2%, CVSS 7.8.
• Rank 9, CVE-2026-43500, kernel on both hosts: EPSS 94.1%, CVSS 7.8.
• Rank 10, CVE-2025-15467, openssl and libssl3t64 on hyp-db-01: EPSS 74.8%, CVSS 8.8.
• Rank 11, CVE-2025-22037, kernel on both hosts: EPSS 68.8%, CVSS 7.5.
• Rank 12, CVE-2026-32746, telnet on hyp-db-01: EPSS 45.3%, CVSS 9.8. Removing telnet
  fixes this one and rank 1 together.


2. High CVSS score only, no exploitation evidence

These are the top Critical results by score. Every one is CVSS 10.0 (version 3.1) in a
kernel image package (source: wazuh_vulnerabilities):

CVE       CVE-2026-74705, CVE-2026-74612, CVE-2026-74475
Hosts     hyp-web-01 (001), hyp-db-01 (002)
Packages  linux-image 6.8.0-88, 6.8.0-90 and 6.8.0-147

CVE       CVE-2026-74279, CVE-2026-72421, CVE-2026-74280, CVE-2026-63795
Hosts     hyp-web-01 (001), hyp-db-01 (002)
Packages  linux-image 6.8.0-88 / 6.8.0-90

Why these rank lower: none is on the KEV list, and none appears in the exploit ranker's
top 25. Their EPSS is below about 1.9%, the lowest score in that top 25. The
descriptions are kernel bugs in networking and crypto code, such as a UDP tunnel
use-after-free, veth XDP length accounting and a cavium DMA cleanup. They are worth
fixing, but nothing shows anyone is exploiting them. There are also older Critical 9.8s
with low EPSS lower in the ranking, such as CVE-2016-9843 in klibc (rank 14, 5.8%) and
CVE-2021-3773 in the kernel (rank 16, 5.3%).


Patch order I recommend

1. hyp-db-01 (002): remove or upgrade telnet and inetutils, which clears ranks 1 and 12.
   Then upgrade libbrotli1 (rank 3, deadline today) and openssl (rank 10).
2. Both hosts: move to a patched kernel, reboot into it, and remove the old 6.8.0-88 and
   6.8.0-90 kernel packages. That clears five of the seven known-exploited CVEs and
   most of the CVSS 10 list. hyp-web-01 is still running 6.8.0-90 (source:
   wazuh_vulnerabilities). Three CVSS 10 CVEs are also flagged on 6.8.0-147, so a newer
   kernel than that is needed to clear them.
3. After that, work through the rest of the ranked list.

Confidence and limits: confidence is high for the order of the items shown. Wazuh found
them using Canonical's security tracker, and the exploit evidence comes from CISA KEV
and EPSS. One gap: only 2,002 of the 8,062 CVEs were ranked. The KEV check is reported
as fully completed, but I can't be sure from the output that it covered all 8,062. I
also haven't checked whether any of these are reachable from the network on these hosts,
so that is not part of the ranking.

Tools used: wazuh_cve_priority, wazuh_vulnerabilities ×3

68.1s · 5 tools · 0 approvals

A real turn on our own install (CLI 0.6.0, 10 Nov 2026), replayed at speed. Names changed. Type the same words into quox.
✔ Machine verified against the CLI on 10 Nov 2026 · test report mwaptv

ARGUS read Wazuh’s vulnerability inventory and joined it with the CISA KEV catalogue and FIRST EPSS scores. A CVSS 10 with no known exploit ranks below a CVSS 7.8 that attackers are already using. It says when a feed is unavailable rather than ranking without it, and it changes nothing on your servers.

exploit-aware ranking
quox wazuh vulns --priority --limit 8Ⓠ QUOX │ wazuh vulns │ 2002 CVEs, 7 in CISA KEV, 10 with EPSS … ⚠⚠ ranked 2002 of 8062 CVEs: the top by CVSS plus every CVE in CISA KEV. Narrow with --agent.RANK  CVE              SEV       CVSS  EPSS   KEV  HOSTS  PACKAGE1     CVE-2026-24061   Critical  9.8   99.0%  yes  1      inetutils-ping, inetutils-telnet2     CVE-2026-31431   High      7.8   98.1%  yes  2      linux-image-6.8.0-90-generic, linux-image-6.8.0-88-generic3     CVE-2026-111497  High      8.1   82.6%  yes  1      libbrotli14     CVE-2026-46300   High      7.8   16.1%  yes  2      linux-image-6.8.0-90-generic, linux-image-6.8.0-88-generic5     CVE-2025-39964   Medium    5.5   1.6%   yes  1      linux-image-6.8.0-88-generic6     CVE-2026-53266   High      7     0.9%   yes  2      linux-image-6.8.0-90-generic, linux-image-6.8.0-88-generic7     CVE-2026-53362   High      7.8   0.7%   yes  2      linux-image-6.8.0-90-generic, linux-image-6.8.0-88-generic8     CVE-2026-43284   High      7.8   94.2%  no   2      linux-image-6.8.0-90-generic, linux-image-6.8.0-88-generic
From a real run on our demo install (two Ubuntu servers).

KEV yes means CISA lists the CVE as exploited in the wild. EPSS is the estimated chance of exploitation in the next 30 days. The first line says how many CVEs were ranked out of the total, so you never mistake a partial ranking for a complete one. --agent <id> narrows it to one server.

quox wazuh vulns lists the raw inventory, highest CVSS first, when you want the full list rather than the ranking.

Quox read Wazuh’s vulnerability data and two public feeds. Reading changes nothing, so no approval was needed. Patching stays with you and your change process.