A working exploit is now public for CVE-2026-115731, a critical remote code execution flaw in OpenVox and Puppet Server, after Vox Pupuli, Perforce and researchers at Neodyme disclosed the bug on November 3. The vulnerability has a CVSS 4.0 score of 9.4, and no exploitation in the wild has been confirmed so far.
How a Crafted Report Gives an Agent Control of the Server
The flaw lets any agent holding a signed certificate run code on the Puppet server by sending a crafted report. Puppet agents routinely send reports to the server after each run, so the attack path uses a normal channel in the architecture. The OpenVox GitHub advisory, tracked as GHSA-ffr6-rwpm-p5wq, includes a working malicious report that runs a shell command. Neodyme, whose researchers took part in the disclosure, published its own technical write-up.
The affected products are OpenVox and Puppet Server, including Puppet Enterprise. Operators are told to update OpenVox or Puppet Enterprise to patched releases.
CVE-2026-115732 Lets One Node Impersonate Another
A second bug, CVE-2026-115732, carries a CVSS score of 9.0 and lets one node impersonate another. The advisory material treats it as a companion to the code execution flaw. Node impersonation in a configuration management system is significant because the server decides what to send to each machine based on the identity it believes the agent holds.
Autosigning Removes the Certificate Barrier
The requirement for a signed certificate looks like a meaningful barrier, but the picture changes with autosigning. Code execution as the puppet user gives access to the certificate authority signing key. With autosigning enabled, which is common in enterprise deployments, an attacker can simply request a certificate and receive one. In that configuration the chain becomes effectively unauthenticated.
Why the CA Signing Key Matters
The signing key is the root of trust for every node the server manages. An attacker holding it can mint certificates that the infrastructure accepts as legitimate. That turns a single server compromise into a durable foothold, because forged identities can persist after the original bug is fixed unless operators rotate trust material.
Blast Radius Covers Every Node a Puppet Server Manages
The central concern is reach. One compromised Puppet server can reach every node it manages. Configuration management servers sit at the center of an environment: they push packages, files and settings to large numbers of machines, often with elevated privileges. An attacker who controls the server can use that same channel to run code on managed hosts.
This explains the high severity rating even though the flaw requires a certificate. In environments where certificates are easy to obtain, the requirement offers little protection. In environments with manual signing, the risk shifts toward any compromised or malicious agent that already holds a valid certificate, such as a machine an attacker took over earlier.
The public proof of concept shortens the timeline defenders have. With a working malicious report available in an advisory, attackers do not need to reverse-engineer the flaw. Coordinated disclosure typically pairs patches with details so that defenders can test, and the same material also serves as an instruction set for anyone who wants to try the attack against unpatched servers.
Patch Priorities for Puppet and OpenVox Operators
Operators of OpenVox and Puppet Enterprise should update to the patched releases. Those with autosigning enabled have the strongest reason to move quickly, since it removes the one barrier that otherwise limits who can exploit the bug. Administrators can also review whether autosigning is needed at all, and check server logs for unusual report submissions from agents.
No source has reported attacks so far. That status describes the first day after disclosure, and a published exploit for a bug that grants control of a management server is the kind of material that attackers tend to test quickly.
