A cybersecurity analyst monitors an active vulnerability alert, patch status, and certificate settings across a network.
CISA has added CVE-2026-115731, a Perforce Puppet Enterprise deserialization flaw, to its Known Exploited Vulnerabilities (KEV) Catalog. The bug lets deserialization of untrusted data in Puppet's report-processing code lead to arbitrary code execution. Puppet is widely used to configure and enforce state on large server and workstation fleets, and that includes Windows nodes. A compromise of the Puppet Server is therefore a much bigger problem than one lost host.

CISA's November 17 alert is short. It confirms the addition "based on evidence of active exploitation". It also calls this class of bug a frequent attack vector that poses significant risk to the federal enterprise. The alert does not say who is exploiting the flaw, where, or how. Treat any claim about specific campaigns as unsupported for now.

A cybersecurity analyst monitors an active vulnerability alert, patch status, and certificate settings across a network. What the vulnerability does​

According to the CVE record, a client holding any CA-signed agent certificate can execute arbitrary code on the Puppet Server as its service user. It does this by submitting a report containing a crafted rich-data type via a PUT request. The weakness is classed as CWE-502, deserialization of untrusted data. Some trackers also list CWE-94, code injection.

The autosigning condition is the key detail. Red Hat's entry says that where the certificate authority autosigns requests without strong verification, an unauthenticated network attacker can obtain such a certificate. Without that setting, an attacker typically needs an already-authorized agent certificate. So this is not automatically a no-credentials, internet-wide bug in every deployment. It also isn't a flaw that only trusted insiders can reach. Your CA policy decides which of the two you are.

Why a Puppet Server compromise is serious​

Red Hat rates the flaw Critical. It says successful exploitation gives remote command execution with the privileges of the Puppet service account in the Puppet Server. It adds that this can compromise the central certificate authority and all downstream managed systems. That is Red Hat's own assessment of potential impact. It is not evidence that those systems were compromised in the exploitation CISA cited.

Think of the Puppet Server as the master keyring for a building. Whoever picks that lock does not need to pick the doors one at a time.

Severity scores​

  • The CNA score is CVSS v4.0 9.4 (Critical). Trackers list 9.4/10 on the CVSS scale.
  • Red Hat lists a separate CVSS v3.1 score of 9.9. It explains that its scores reflect how a flaw affects its own products, so they can differ from NVD or other vendors.
  • These are different scoring versions and contexts, not contradictory measurements.
  • Perforce's CVE dashboard lists the CVE as Critical with a 9.4 base score and a resolution date of 11/3/2026.

Affected and fixed versions​

Perforce's CVE dashboard lists these fixed releases, with the bundled Puppet Core agent version in parentheses:

Product lineFixed release
Puppet Enterprise 2026.02026.0.1 (Puppet Core 9.1.3)
Puppet Enterprise 2025.122025.12.1 (Puppet Core 8.22.2)
Puppet Enterprise 2023.82023.8.13 (Puppet Core 8.22.2)
Puppet Core9.1.3 and 8.22.2

One third-party aggregator lists the affected Puppet Enterprise ranges as 2023.8.0 up to but excluding 2023.8.13, 2025.0.0 up to but excluding 2025.12.1, and 2026.0.0 up to but excluding 2026.0.1. Treat that as a working inventory guide and confirm it against Perforce's own advisory.

One caution: another tracker wrote the affected releases as if the fixed versions themselves were vulnerable. The version boundaries above are the better reading. The fixed releases are the safe ones.

The same dashboard lists a second Critical Puppet flaw fixed in those same releases. CVE-2026-115732 is a Certificate Name Authorization Bypass with a 9.0 score. I found no indication that CISA has listed that one. If you patch for one, you cover both, so it is worth knowing about.

Open-source forks and distribution packages are affected as well. Red Hat's entry mentions Satellite environments and points to OpenVox. Its AI-generated mitigation note names OpenVox Server 9.2.0 and 8.18.0 as fixed builds. Red Hat labels that guidance as a courtesy and tells customers to test it themselves, so confirm those versions against the OpenVox advisory. FreeBSD and Tenable Nessus have also published entries for OpenVox and RHEL 9 Satellite 6.19.

Timeline​

  • November 3, 2026: the CVE was published and Perforce released fixes. Debian lists the same publication date.
  • November 17, 2026: CISA adds it to KEV.

That is two weeks between disclosure and the catalog listing. Administrators who treated this as a routine monthly patch have probably already lost some of that window. Perforce's initial notice reportedly said it knew of no active exploitation. That was a statement about November 3, and it does not conflict with CISA's later finding. I could not retrieve Perforce's Trust Center page directly, so I am not quoting its wording.

What to do now​

  1. Inventory. Find every Puppet Enterprise, Puppet Core and OpenVox server, including lab, staging and forgotten ones. Record the version of each.
  2. Upgrade. Move to the fixed release for your line (2023.8.13, 2025.12.1 or 2026.0.1 for Puppet Enterprise). Restart the server so the fix loads, as Perforce advises.
  3. Review autosigning. Red Hat's interim mitigation is to avoid autosigning certificate requests without strong verification. Blanket autosign settings are the biggest multiplier of exposure here.
  4. Limit network reach. Red Hat's mitigation note suggests restricting access to the report endpoint (PUT /puppet/v3/report). Trackers also advise restricting access to Puppet Server generally. This reduces exposure but does not replace patching.
  5. Hunt for compromise. Because exploitation is confirmed, patching alone is not the end of the job. Red Hat's note suggests EDR and auditd coverage for unusual child processes or command executions from the puppetserver service user, and runtime tools such as Falco. Check recently signed certificates and unexpected agent registrations. Preserve logs before you rebuild anything. This is prudent practice drawn from that vendor guidance, not a checklist CISA published.
  6. Consider rotation. If you find signs of compromise, treat the CA and any secrets the server handled as exposed, and plan to rotate them. That is general incident-response practice, not something the sources prescribe.

What KEV status means​

The CISA alert ties the listing to Binding Operational Directive 26-04, "Prioritizing Security Updates Based on Risk". The alert says the directive applies to Federal Civilian Executive Branch agencies. It tells them to prioritize fast remediation of KEV-listed CVEs on publicly exposed assets that grant total control after exploitation. It also sets expectations for checking whether a system was compromised before it was patched. I could not retrieve the directive's full text, so I won't state deadlines or specific forensic requirements.

For everyone else, the alert says that BOD 26-04 applies only to FCEB agencies, while CISA encourages all organizations to adopt risk-based vulnerability management. That is encouragement, not a legal obligation. Still, KEV is a handy tiebreaker when your patch queue is longer than your maintenance window. A bug that attackers are known to exploit, in a tool that holds the keys to your fleet, goes near the top of the queue.

Bottom line​

This is not a Windows-specific bug. Windows shops that use Puppet to manage Windows agents, though, are exactly the shops that depend on the Puppet Server being trustworthy. The practical sequence is to patch to the fixed release, restart the server, lock down autosigning, and then look for signs the server was hit before you got there. The public reporting does not say how widespread exploitation is, so assume your exposure depends on your certificate policy and network layout, and verify both.

 

References

  1. CISA Adds One Known Exploited Vulnerability to Catalog CISA 2026-11-17T12:00:00+00:00
  2. CVE-2026-115731 - Red Hat Customer Portal access.redhat.com
  3. CVE-2026-115731 Report Deserialization Code Execution - vulnerability database vulners.com