2026-11-21
MONAI Label Critical Path Traversal Flaw (No CVE, 0.8.5 and Earlier) Reported Exploited With an Autonomous AI Tool: What Server Operators Should Do
On 18 November 2026 the MONAI Label project published a GitHub security advisory (GHSA-qc22-rmgc-vg94, rated Critical, CWE-22, no CVE assigned) for versions 0.8.5 and earlier of its medical-image annotation server. Two datastore upload endpoints use a caller-supplied ID without checking the resulting file path, so a client can write files outside the intended directory. Authentication is disabled by default, so in the default configuration any client that can reach the server can use it; with authentication on, the annotator role is enough. The advisory lists no patched version: the fix is in a pull request (#1915) that the project merged to its main branch on 19 October 2026 and says will ship in the next release. As of 21 November 2026 we found no release newer than 0.8.5 on PyPI. BleepingComputer reports that a cloud provider's threat intelligence team observed attackers using an autonomous AI penetration-testing platform to find this flaw and take over servers, mainly at education institutions in East Asia, on 8 to 10 September 2026. This site's take: do not expose MONAI Label to the internet, turn authentication on, restrict who can reach it, and move to a fixed build when one is released.