
Cloud Vulnerability DB
A community-led vulnerabilities database
CVE-2026-121520 is a privilege escalation vulnerability in Moby BuildKit that allows a malicious client with access to the BuildKit control APIs to set extended attributes (xattrs) on files outside the BuildKit state directory. Affected versions are all BuildKit releases prior to v0.34.1 (i.e., <= v0.34.0). The vulnerability was published on November 3, 2026, and fixed in v0.34.1. It carries a CVSS v3.1 base score of 7.9 (High) (GitHub Advisory, Red Hat Bugzilla).
The root cause is improper path restriction (CWE-22: Path Traversal) combined with symlink-following behavior (CWE-61: UNIX Symbolic Link Following), which allows a crafted build request to escape the intended BuildKit state directory boundary when setting extended file attributes. An attacker with access to the BuildKit control APIs submits a malicious build request that specifies an xattr write targeting a file path outside the BuildKit state directory. Critically, xattrs in the trusted. or security. namespaces can be set this way, which may grant elevated privileges to programs on the host system. The attack vector is local and requires high privileges (API access), but the scope change means impact extends beyond the BuildKit process itself (GitHub Advisory, Red Hat Bugzilla).
A successful exploit allows an authenticated malicious client to write arbitrary extended attributes to files anywhere on the host filesystem, bypassing the intended containment of the BuildKit state directory. By setting xattrs in the trusted. or security. namespaces (e.g., Linux capabilities or SELinux labels), an attacker could elevate the privileges of a host program beyond what it would normally have, potentially achieving full host compromise. There is no confidentiality impact, but integrity and availability of the host system are both rated High in the CVSS v3.1 assessment (GitHub Advisory, Red Hat Bugzilla).
There is no public proof-of-concept exploit and no evidence of in-the-wild exploitation at this time. The EPSS score is approximately 0.139%, indicating a low probability of exploitation in the near term. The vulnerability is not listed in the CISA Known Exploited Vulnerabilities (KEV) catalog. Exploitation is constrained by the requirement that the attacker already have access to the BuildKit control APIs, limiting the attack surface to multi-tenant or shared build service environments (GitHub Advisory, Feedly).
trusted. or security. namespace (e.g., security.capability) to assign Linux capabilities or other privilege-elevating attributes to a target binary on the host.../ sequences or absolute paths outside the expected state directory (e.g., /var/lib/buildkit).getfattr -d <file> — particularly attributes in the trusted.* or security.* namespaces on files not normally associated with BuildKit.The fix is available in BuildKit v0.34.1; all users should upgrade immediately. As a workaround, restrict access to the BuildKit control APIs exclusively to trusted users and processes — the vulnerability only affects deployments where untrusted users can submit build requests. Additionally, implementing network segmentation to limit which systems can reach the BuildKit service reduces exposure. Red Hat has tracked this issue for affected downstream packages (GitHub Advisory, Red Hat Bugzilla).
The advisory was published by BuildKit maintainer tonistiigi on GitHub with a "Moderate" severity rating. Red Hat's Product Security team tracked the issue as high priority/severity in their Bugzilla system, with over 120 engineers CC'd, indicating broad downstream concern across Red Hat products. No significant public social media discussion or independent researcher commentary has been identified at this time (GitHub Advisory, Red Hat Bugzilla).
Fix availability across major Linux distributions and their releases.
bionic (esm-apps)
docker.io
devel
docker.io
focal (esm-apps)
docker.io
jammy
docker.io
jammy (esm-apps)
docker.io
noble
docker.io
noble (esm-apps)
docker.io
resolute
docker.io
Source: This report was generated using AI
Free Vulnerability Assessment
Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.
Get a personalized demo
"Best User Experience I have ever seen, provides full visibility to cloud workloads."
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
"We know that if Wiz identifies something as critical, it actually is."