Critical MONAI Label file-write flaw has no patched package release

MONAI Label monailabel <= 0.8.5 has a Critical arbitrary-file-write flaw with observed exploitation and no patched package release named at the reporting cutoff.

Reader ratingRate this article

Loading reader rating…

Privacy-conscious voting uses a signed random browser identifier and a short-lived network hash; the rating tables contain no raw IP-address column.

Reader ratings require JavaScript; the article remains fully readable without it.

Security advisory • reporting cutoff 20 November 2026, 1:33 p.m. EST (18:33 UTC)

Critical MONAI Label file-write flaw has no patched package release [1]

GitHub lists monailabel <= 0.8.5 as affected and patched versions as “None.” [1]

AWS reports successful arbitrary file writing and a later root reverse shell in its observed campaign. [2]

Bottom line: A reachable MONAI Label server running monailabel <= 0.8.5 needs immediate containment because no patched package release was named at the cutoff. [1]

Status at the cutoff

Affected monailabel <= 0.8.5 [1]
Package fix Patched versions: None. The fix is merged in PR #1915, but the advisory says it will be included in the next MONAI Label release; it does not name the first fixed package version. [1]
Identifiers GitHub lists No known CVE and CWE-22. No CVSS score appears in the frozen advisory capture. [1]
Severity GitHub rates the advisory Critical. [1]

Who is exposed?

The exposure predicate is conjunctive: an affected monailabel <= 0.8.5 server is running, an untrusted client can reach it, and authentication is disabled by default. [1]

If authentication is enabled, access with the ANNOTATOR role is sufficient. [1]

The file-write impact extends only to locations where the MONAI Label process has operating-system write permission. [1]

What AWS observed

AWS says its observed agents moved from mapping the API to arbitrary file writing in about 11 minutes. [2]

AWS reports that a later session used the same flaw to establish a reverse shell running as root. [2]

AWS assesses a Chinese nexus with moderate confidence. [2]

AWS says it inferred that the second university server had been compromised to establish a tunnel. [2]

Prioritized defensive guidance

  1. Stop untrusted reachability. The maintainer says not to expose the MONAI Label server to untrusted networks. [1]
  2. Restrict application access. Enable authentication and do not grant ANNOTATOR to untrusted users. [1]
  3. Reduce process privilege. Run the server as an unprivileged user with write access limited to the datastore directory. [1]
  4. Track the packaged fix. PR #1915 is merged, but the first fixed package release is not named; verify the installed package version after a release is published. [1]

Verify the outcome

After changes, verify authentication is enabled, untrusted users lack the ANNOTATOR role, the server is not exposed to untrusted networks, and the service runs unprivileged with write access limited to the datastore. [1]

Local relevance

Local relevance: Treasure Coast healthcare, research, imaging, and AI teams need to act only if their environment matches the exposure predicate above; this article does not establish that any local organization runs MONAI Label. [1]

What remains unknown

  • The first fixed package version and release date remain unknown at the cutoff. [1]
  • Whether a CVE or CVSS score will be assigned remains unknown. [1]

Sources

  1. GitHub Security Advisory GHSA-qc22-rmgc-vg94 — Project MONAI/MONAI Label; live capture accessed 20 November 2026, 1:33 p.m. EST (18:33 UTC).
  2. AWS Security Blog: Chinese-nexus actor’s Cairn AI agents found a zero-day on their own — published 20 November 2026; live capture accessed 20 November 2026, 1:33 p.m. EST (18:33 UTC).