Skip to content
← Blog · · df00tech

CVE-2026-115731: Critical Deserialization RCE in Perforce Puppet Enterprise (CISA KEV)

vuln-intel Perforce puppet CVE-2026-115731 #checker-reject

What the vulnerability is

CVE-2026-115731 is a critical (CVSS 9.4) deserialization of untrusted data vulnerability (CWE-502/CWE-94) affecting Puppet, as used by Puppet Server and Puppet Enterprise. A client holding any CA-signed agent certificate can achieve arbitrary code execution on the Puppet Server, running as the service user, by submitting a report containing a crafted rich-data type via a PUT request to the report endpoint. The root cause is unsafe handling of rich-data types during report deserialization, allowing an attacker-controlled object graph to trigger code execution on the server side.

The barrier to exploitation is lower than it first appears: where the certificate authority autosigns agent certificate requests without strong verification, an unauthenticated network attacker can simply request and obtain a valid CA-signed certificate, then use it as the credential needed to reach the vulnerable endpoint.

Affected software

The vulnerability affects Puppet / Puppet Enterprise from Perforce:

  • Puppet / Puppet Enterprise >= 8.0.0, < 8.22.2
  • Puppet / Puppet Enterprise >= 9.0.0, < 9.1.3
  • Puppet Enterprise >= 2023.8.0, < 2023.8.13
  • Puppet Enterprise >= 2025.0.0, < 2025.12.1
  • Puppet Enterprise 2026.0.0

Exploitation status

This vulnerability is listed in the CISA Known Exploited Vulnerabilities (KEV) catalog, meaning it has been observed exploited in the wild. Defenders should treat any internet-reachable or improperly segmented Puppet Server — particularly one with autosigning enabled — as an active target, not a theoretical risk, and prioritize patching and detection accordingly.

How our detection catches it

The rule is built around the mechanics of this specific exploitation chain rather than a single IOC, and ships across Microsoft Sentinel (KQL), Splunk (SPL), Elastic (EQL), QRadar (AQL), Sumo Logic, Chronicle (YARA-L), and CrowdStrike (CQL). At a high level, the logic looks for:

  • Anomalous child processes spawned by Puppet or JRuby server processes — the telltale sign of code execution reached through the deserialization path rather than normal Puppet catalog/report operations.
  • Suspicious inbound requests to the Puppet report endpoint, flagging PUT requests with characteristics consistent with crafted rich-data payloads rather than legitimate agent report submissions.
  • Unexpected outbound network connections originating from the Puppet infrastructure tier, which can indicate post-exploitation callback or staging activity from a compromised Puppet Server.

Correlating these three signals — process lineage, endpoint-specific request anomalies, and outbound network behavior from Puppet infrastructure — gives defenders coverage even if any single signal alone would be too noisy to act on.

Remediation and further reading

Organizations running affected Puppet Server or Puppet Enterprise versions should patch immediately given the KEV listing, and review certificate autosigning configuration to eliminate the unauthenticated path to a valid agent certificate. For the full detection logic, sample events, and per-SIEM queries, see the CVE-2026-115731 detection page.

Get new detections in your inbox

New ATT&CK coverage plus CISA KEV / CVE detection rules, roughly weekly. No spam, unsubscribe anytime.