CVE-2026-115731: Puppet Server RCE
Active exploitation confirmed in CISA KEV.
Active exploitation confirmed in the wild. CISA added this to the KEV catalog on 2026-11-17. Federal agencies must patch by 2026-12-01.
TL;DR - CVE-2026-115731 is a critical unsafe-deserialization RCE in Puppet Server report processing, rated 9.9 by NVD. - Puppet Enterprise versions before 2023.8.13, 2025.12.1, and 2026.0.1 are affected. - CISA added the CVE to its Known Exploited Vulnerabilities catalog on November 17, 2026, confirming exploitation in the wild. - Upgrade urgently, review certificate autosigning, restrict port
8140, and investigate suspicious report submissions.
What Happened?
CVE-2026-115731 is a remote-code-execution vulnerability in the report-processing code used by Puppet Server and Puppet Enterprise. A client with a CA-signed Puppet agent certificate can submit a specially crafted report through a PUT request. When the server processes attacker-controlled rich-data content, unsafe deserialization can execute arbitrary code as the Puppet Server service account.
The Puppet Server vulnerability becomes more serious when certificate requests are automatically signed without strong identity validation. In that configuration, an unauthenticated network attacker may be able to obtain a CA-signed agent certificate and then submit a malicious report.
| Field | Value |
|---|---|
| CVE | CVE-2026-115731 |
| Severity | Critical |
| NVD CVSS base score | 9.9 |
| Attack vector | Network |
| Certificate requirement | A CA-signed agent certificate is generally required; weak autosigning may enable certificate acquisition without strong authentication |
| Privileges required | No conventional application privilege is required once a usable certificate is obtained |
| Patch available | Yes |
| CISA KEV status | Added November 17, 2026 |
| Federal remediation due date | December 1, 2026 |
| Known ransomware use | Unknown |
NVD published the vulnerability on November 3, 2026, and last modified the record on November 17, 2026. The available research includes the NVD score but not the complete CVSS vector string. Verify the vector directly in the NVD CVE-2026-115731 record before using individual metric values for formal risk calculations.
Analyst’s take: Identify exposed Puppet control planes first. Then determine whether certificate autosigning could provide an unauthenticated path to a usable agent certificate. KEV-listed exploitation makes patching urgent, while network restriction and certificate-policy review can narrow the attack path during the upgrade window.
Root Cause and Attack Path
The root cause is unsafe deserialization of attacker-controlled data in Puppet’s report-processing path. Puppet reports can contain rich-data types. In vulnerable releases, a crafted type supplied in a report can reach deserialization logic that does not sufficiently restrict which objects or types may be reconstructed. The resulting behavior permits attacker-controlled code execution on the Puppet Server.
The vulnerable operation occurs on the server after it receives a report, making the Puppet control plane the primary target. Successful execution occurs under the Puppet Server service account, but the practical impact may extend beyond that account because Puppet servers commonly hold credentials, signed certificates, configuration data, and administrative connectivity to managed infrastructure.
The certificate requirement is central to the threat model. A valid CA-signed agent certificate establishes enough trust for an attacker to submit the relevant report. Deployments using strict manual approval and identity verification reduce the likelihood of certificate acquisition. Deployments that autosign arbitrary or weakly validated requests can expose a chain of attack:
- Obtain a usable agent certificate.
- Submit a crafted Puppet report.
- Trigger unsafe deserialization.
- Execute code on the Puppet Server.
Technical Notes
The relevant request is a report submission to the Puppet Server endpoint, normally served over HTTPS on port 8140. The exact URI and request structure vary by Puppet deployment and version, so defenders should not rely on a single path-based indicator.
A conceptual malicious request has this general form:
PUT /<puppet-report-endpoint> HTTP/1.1
Host: puppet.example.com:8140
Content-Type: application/json
Authorization: <client-certificate-authentication>
Content-Length: <length>
<crafted Puppet report containing malicious rich-data type>
Do not reproduce or test exploit payloads against production systems. Use vendor guidance and an isolated validation environment when confirming exposure.
Affected Versions and Fixed Releases
The directly identified affected product is Perforce Puppet Enterprise.
| Product | Affected versions | Fixed version |
|---|---|---|
| Puppet Enterprise | 2023.8.0 through versions before 2023.8.13 | 2023.8.13 |
| Puppet Enterprise | 2025.0.0 through versions before 2025.12.1 | 2025.12.1 |
| Puppet Enterprise | 2026.0.0 through versions before 2026.0.1 | 2026.0.1 |
Organizations should identify the active release branch before selecting a package. Upgrade to the fixed release corresponding to that branch rather than assuming that a later release from a different branch is automatically supported by the existing control-plane topology, modules, operating system, or operational tooling.
Related OpenVox and Puppet component version information has appeared in secondary research results, but the retrieved Perforce advisory material did not independently verify all reported ranges. Reported OpenVox fixes include OpenVox Server 8.18.0 and 9.2.0, OpenVox Agent 8.32.0 and 9.3.0, and Puppet Core fixes at 8.22.2 and 9.1.3. Treat these as leads for inventory and validation, not authoritative patch instructions, until confirmed in the relevant official project advisory.
Inventory Commands
Start with an inventory of server packages, release manifests, and service endpoints:
puppetserver --version
puppet --version
rpm -qa | grep -Ei 'puppet|openvox'
dpkg-query -W | grep -Ei 'puppet|openvox'
Identify listeners and reverse proxies exposing the Puppet service:
ss -lntp | grep ':8140'
Record whether the host is Puppet Enterprise, standalone Puppet Server, or an OpenVox-based deployment. Package names alone may not identify the control-plane release accurately, so confirm the version through the platform’s supported administration interface or release documentation.
Why Is the CVSS Score Critical?
NVD assigns CVE-2026-115731 a CVSS base score of 9.9, placing it in the critical severity range. The score reflects network reachability, remote code execution, and the potential impact on a centralized configuration-management server.
The vulnerability can be triggered through a network-accessible report submission path. It does not require the attacker to log in as a conventional administrative user. Instead, the attacker needs a CA-signed agent certificate or must exploit a weak certificate-autosigning configuration to obtain one.
A high score does not mean every installation is unauthenticated or internet exploitable. A deployment with private network access, strict certificate approval, and strong monitoring may have a narrower attack path. Those controls do not eliminate the underlying vulnerability, and CISA KEV status means defenders should not defer remediation while waiting for evidence of local targeting.
The exact NVD vector string was not included in the supplied record, so individual CVSS metric labels should not be inferred beyond the documented behavior. Organizations should supplement CVSS with exposure, certificate-policy, network-reachability, and control-plane-criticality assessments.
Has CVE-2026-115731 Been Exploited?
Exploitation in the wild is confirmed. CISA added CVE-2026-115731 to its Known Exploited Vulnerabilities catalog on November 17, 2026. The catalog entry sets a federal remediation due date of December 1, 2026.
KEV inclusion means this should be treated as an active exploitation risk rather than a theoretical vulnerability or a finding based solely on a laboratory proof of concept.
CISA’s record does not identify confirmed ransomware-campaign use. The known-ransomware-campaign field is Unknown. That does not indicate safety; it only means the available CISA classification does not confirm that ransomware groups have used this CVE.
A public proof of concept is also reported in secondary search results, including references to an OpenVox GitHub security advisory and a sample malicious report. The retrieved research did not provide a directly verified advisory URL or independently validate the repository contents. Defenders should therefore describe the PoC as reported but requiring direct confirmation. The confirmed fact is KEV-listed exploitation.
Do not download or execute unverified exploit code on production Puppet infrastructure.
Detection Guidance
Detection should begin with exposure assessment:
- Identify Puppet Server and Puppet Enterprise hosts.
- Determine whether port
8140is reachable from untrusted networks. - Review the certificate authority’s signing policy.
- Identify automatic-signing rules that accept broad certificate names, wildcard patterns, unverified hostnames, or requests from untrusted networks.
- Review Puppet Server, proxy, certificate-authority, and host telemetry together.
Review Puppet Server access logs for unusual report submissions, especially from newly issued certificates, unfamiliar source addresses, unexpected agent names, or bursts of PUT requests. Log locations vary by distribution and deployment architecture, so inspect both Puppet Server logs and any reverse-proxy or load-balancer logs.
Log Review
A preliminary search for suspicious report activity can begin with web and proxy logs:
grep -E '"PUT [^"]*(report|catalog|facts)' /var/log/puppetlabs/puppetserver/* \
/var/log/nginx/* /var/log/httpd/* 2>/dev/null
The command is only a starting point. Validate the actual log format and endpoint in the environment.
Useful investigation fields include:
- Timestamp and source IP for report
PUTrequests - Client certificate subject and serial number
- Agent certificate issuance time
- Agent name and certificate common name
- Request size and content type
- HTTP status code and response latency
- Report-processing exceptions or deserialization errors
- Child processes launched by the Puppet Server service account
A generic SIEM query can correlate report submissions with certificate issuance:
service="puppetserver"
http_method="PUT"
(uri contains "report" OR request_body contains "rich-data")
| correlate client_certificate_serial
| join certificate_authority_events where action="sign"
| where certificate_age < 24h OR source_ip not in approved_puppet_networks
Adapt field names to the organization’s SIEM. A strong detection signal is a newly issued agent certificate followed shortly by a report submission from an unusual address and then process creation on the Puppet Server host.
On Linux, review process telemetry for unexpected commands launched by the Puppet Server account:
ausearch -m EXECVE -ts recent | grep -Ei 'puppet|puppetserver'
journalctl --since "24 hours ago" _COMM=java
These commands do not prove exploitation. Investigate any matches alongside Puppet, certificate-authority, proxy, and host telemetry.
Remediation and Patching
Upgrade Puppet Enterprise to the fixed version for the deployed release branch:
- Upgrade
2023.8.xinstallations to 2023.8.13 or a later supported release. - Upgrade
2025.xinstallations to 2025.12.1 or a later supported release. - Upgrade
2026.0.xinstallations to 2026.0.1 or a later supported release.
Use the Perforce-supported upgrade workflow for the specific Puppet Enterprise topology. Do not replace packages manually unless that method is supported for the installation.
Before changing the control plane, confirm:
- Backups and restoration procedures
- Database and configuration dependencies
- Module compatibility
- Certificate availability
- Operating-system support
- Rollback procedures
- The resulting version of every relevant control-plane component
Until the upgrade is complete:
- Restrict access to the Puppet Server report endpoint.
- Limit port
8140to known Puppet agents, administration networks, and approved proxies. - Disable automatic certificate signing where operationally possible.
- Narrow autosigning rules if autosigning must remain enabled.
- Require strong verification of agent identity before signing requests.
- Monitor certificate issuance and report submissions for anomalies.
Package and Version Checks
The exact upgrade command depends on the installation method and operating system. First inspect configured repositories and available versions:
rpm -qa | grep -Ei 'puppetserver|puppet-agent'
dnf list --showduplicates puppetserver puppet-agent 2>/dev/null
apt-cache policy puppetserver puppet-agent 2>/dev/null
After selecting the vendor-supported fixed package and completing change approval, the package operation may resemble:
# Use only the repository and package versions documented for your deployment
sudo dnf upgrade puppetserver puppet-agent
# or
sudo apt-get update && sudo apt-get install --only-upgrade puppetserver puppet-agent
For Puppet Enterprise, follow the corresponding Perforce upgrade procedure and verify that the resulting platform reports 2023.8.13, 2025.12.1, or 2026.0.1, as applicable. Do not treat a successful package transaction as proof that every control-plane component has been upgraded.
As a temporary network control, restrict port 8140 at the firewall or security group. An environment-specific rule might allow only the Puppet agent network:
sudo firewall-cmd --permanent \
--add-rich-rule='rule family="ipv4" source address="10.20.0.0/16" port port="8140" protocol="tcp" accept'
sudo firewall-cmd --reload
Replace the example network with the actual approved agent ranges. A firewall rule is a compensating control, not a fix.
If a vulnerable, internet-exposed server cannot be patched or adequately isolated, escalate it as an incident-response priority. CISA guidance states that, where mitigations are unavailable, applicable guidance may require discontinuing use of the product.
Incident-Response Considerations
Treat a vulnerable Puppet Server as a high-value control-plane asset. If you identify suspicious certificate issuance, anomalous report submissions, unexpected child processes, or unexplained configuration changes:
- Preserve Puppet Server, proxy, CA, and host logs.
- Record certificate serial numbers, subjects, issuance times, and revocation status.
- Isolate the server from unnecessary network paths while preserving approved administrative access.
- Investigate processes and files created by the Puppet Server service account.
- Review Puppet manifests, modules, credentials, and downstream system changes for tampering.
- Rotate credentials and certificates that may have been accessible from the server.
- Rebuild the server from a trusted source if compromise cannot be ruled out.
- Validate every connected control-plane and managed-system trust relationship before restoring normal operations.
Do not assume that patching alone removes persistence or reverses changes made during exploitation.
References
-
NVD: CVE-2026-115731
NVD record containing the vulnerability description and CVSS base score. -
Perforce security advisory
Vendor reference for the Puppet report-deserialization issue and fixed release lines. The advisory may require access through the Perforce Support Portal. -
CISA Known Exploited Vulnerabilities Catalog
Confirms KEV inclusion on November 17, 2026, the December 1, 2026, remediation due date, and unknown ransomware-campaign status. -
Red Hat CVE reference
Additional reference describing the report-event deserialization issue and related fixed-build guidance.
Because secondary references contain unverified details for some OpenVox and Puppet Core release lines, organizations should confirm those versions against current official project advisories before applying component-specific remediation. The directly identified Puppet Enterprise fixes are 2023.8.13, 2025.12.1, and 2026.0.1.
For additional security planning, see the web application hardening checklist and the software bill of materials explanation.
This article may contain affiliate links. We earn a commission on qualifying purchases at no extra cost to you.