On this page (6 sections)
CVE-2026-130784 is a critical authorization bypass in Incus, the container and virtual machine manager, affecting versions before 7.6.0. According to the NVD record, a restricted client can create and access default project resources by exploiting entitlement checks against the requested project, and can use that to execute code as root on the host.
If you run Incus on a shared host, give restricted access to other users or teams, or expose its API beyond administrators, treat this as a priority upgrade. The research available so far comes from a single source, so details beyond the CVE description are limited.
What happened
The CVE record, published on 2026-11-18, describes an authorization bypass in Incus. Restricted clients are meant to be limited in what they can reach. Per the description, they can create and access default project resources by exploiting entitlement checks against the requested project.
The description names two consequences. First, an attacker can create bridge networks with raw.dnsmasq directives to execute code as root on the host. Second, an attacker can manipulate inherited resources through the ?project= parameter. The vendor advisory is referenced as GHSA-5c64-j5m6-8433; read it for the authoritative fix details.
Why this class of bug is serious
Authorization bugs (CWE-863) occur when software performs a permission check that does not correctly enforce what a user is allowed to do. In this case a client that is supposed to be restricted can reach resources in the default project that it should not be able to create or use.
The most serious outcome in the CVE description is code execution as root on the host, reached by creating bridge networks with raw.dnsmasq directives. Root on the host is the highest level of access on that machine. The research does not describe further impact beyond this, so any wider consequences, such as effects on instances running on the host, would be an inference and should be assessed by the host owner.
Who is affected
- Hosts running Incus versions before 7.6.0.
- Setups that use restricted clients, for example identities limited to certain projects for tenants, teams or CI jobs.
- Hosts where the Incus API is reachable by anyone other than fully trusted administrators.
- Environments where the default project holds resources that other projects inherit, since the description mentions inherited resources being manipulated.
A single-user workstation where only you hold full admin access has much lower practical exposure, though upgrading is still sensible.
How to check your exposure
First confirm which version each host runs. This read-only command prints the client and server versions; run it on each Incus host. Anything below 7.6.0 is affected.
incus versionNext, list who can talk to the server. This read-only command lists trusted clients and certificates; look for restricted entries you do not recognize or no longer need.
incus config trust listThen review networks for settings that should not be there. These read-only commands list networks and show one network's configuration; replace NETWORK_NAME with your own and look for a raw.dnsmasq key you did not set.
incus network list
incus network show NETWORK_NAMEWhat to do now
- Upgrade Incus to 7.6.0 or later on every host, using your distribution's or vendor's package channel.
- Read the vendor advisory for any additional upgrade notes.
- Until you can upgrade, remove or suspend restricted client access you do not strictly need, and limit network access to the Incus API to trusted addresses.
- Review bridge networks for unexpected raw.dnsmasq settings, and check for resources in the default project that you did not create.
- Audit the trust list and remove stale or unknown client certificates.
- If you find unexpected networks or settings on a host that had restricted clients, treat the host as potentially compromised and follow an incident process, including rotating credentials that were present on it.
After upgrading, run the version command again to confirm 7.6.0 or later is active on each host.
Detection and monitoring
The research gives no indicators of compromise, so rely on general hygiene. Watch for new or changed networks, especially any with raw.dnsmasq keys, and for API requests from restricted clients that use unusual project parameters. If daemon logs are shipped to a central place, keep them for review after an upgrade.
Host-level signals matter too: unexpected dnsmasq configuration, and new files or users on the host. For general guidance on responding to a suspected compromise, see website hacked: what to do.
Incus is infrastructure, but it often sits underneath web workloads. If you host sites on it, Webforta Monitor can alert you to uptime, certificate, DNS and header changes on those sites, which can help you notice trouble sooner.
Sources
This article is based on the public sources below. Details can change as vendors and researchers publish more; check the original advisories before acting on version numbers.
