MLflow 2.x: Actively Exploited, Fixed Only in 3.15 — and Never Declared End of Life

By Scott Bissett Published: August 20, 2026 CISA KEV — federal deadline September 2, 2026
Sep 2, 2026
CISA KEV remediation deadline for CVE-2026-64849
days remaining

MLflow carries an actively exploited, unauthenticated SSRF — CVE-2026-64849, CVSS 9.3 critical. CISA added it to the Known Exploited Vulnerabilities catalog on 2026-08-19; scanning began within hours of CVE assignment. Attackers use it to reach internal and cloud metadata services — the classic path to stolen cloud credentials. The fix exists only in MLflow 3.15.0 and later. If you run any 2.x release, there is no patch for you and there never will be.

Key Dates at a Glance

Running MLflow past end of life?
Extended support past the official EOL date exists for many products in this position — whether it covers MLflow is exactly what we check. Tell us where to reach you and we’ll reply with matched options and pricing guidance — or an honest “no vendor covers this.” Free, no obligation.

Free · No obligation · Independent — we track the dates, vendors don’t pay for placement · dates verified against vendor sources. See all support options →

What the vulnerability actually is

CVE-2026-64849 is a server-side request forgery in MLflow’s webhook delivery. The URL validation in _validate_webhook_url can be bypassed with unvalidated HTTP redirects or DNS rebinding, letting an unauthenticated attacker make the MLflow tracking server fetch arbitrary internal URLs and read the responses. Pointed at a cloud provider’s metadata endpoint, that reads out the instance’s temporary credentials — which is why CISA’s catalog entry describes attackers reaching internal or cloud metadata services. GitHub’s advisory (GHSA-7gwp-5pfp-969j) rates it CVSS 9.3, critical.

One discrepancy worth naming: MLflow’s own advisory does not state a patched version. The GitHub Security Advisory database — the authoritative record for package ecosystems — lists the vulnerable range as everything below 3.15.0 and the first patched version as 3.15.0. That range includes every 2.x and 1.x release ever shipped.

MLflow 2.x is end of life — nobody announced it

MLflow publishes no lifecycle policy: no support windows, no EOL dates, no security-maintenance commitments. Until this week that was an academic complaint. Now it is operational: the 2.x line — the default MLflow for three years, still pinned in thousands of requirements.txt files — has been silent since 2026-05-12, received no backport of an actively exploited critical fix, and got no announcement that support had ended. That is end of life in every way that matters, declared by silence.

LineFirst releaseStatusYour move
MLflow 3.x2025-06-10Active — all fixes land hereBe on 3.15.0 or later. Anything earlier in 3.x is vulnerable to CVE-2026-64849.
MLflow 2.x2022-11-15De facto EOL — no release since 2.22.5Migrate to 3.15.0+. No patch exists for 2.x and none is coming.
MLflow 1.x2019-06-04EOL — last release 2023-04-05Same, with a longer path.

The pattern: third AI infrastructure tool on KEV in six weeks

MLflow is not an outlier. It is the third AI infrastructure tool to land in CISA’s actively-exploited catalog since early July — and none of the three publishes an end-of-life policy:

ToolCVEKEV addedFixed in
MLflow (ML experiment tracking)CVE-2026-64849 — SSRF, CVSS 9.32026-08-19 (due 2026-09-02)3.15.0 only — no 2.x backport
Ray (distributed compute)CVE-2025-62593 — code injection / RCE2026-08-17 (due 2026-08-20)2.52.0, released 2025-11-21 — the fix existed nine months before exploitation forced it onto KEV
Langflow (LLM app builder)Six KEV entries since 2025-05-05Most recent 2026-08-04Per-CVE — run the newest release, always

The common thread: ML tooling runs with cloud credentials, sits on networks its operators think of as internal, and follows a release culture where old versions simply stop existing rather than being formally retired. Platform teams inherit a patching burden the projects themselves never promised to carry.

What to do this week

1. Find your MLflow versions. pip show mlflow per environment, or search your dependency manifests for mlflow pins. Anything below 3.15.0 is vulnerable to an actively exploited critical CVE.

2. Get tracking servers off the open internet. The flaw is unauthenticated. An MLflow tracking server reachable from outside your network is currently being scanned for.

3. Upgrade to 3.15.0 or later. For 3.x deployments that is routine. For 2.x deployments it is a major-version migration — and unlike PHP or Java, no extended-support vendor we track offers post-EOL patching for MLflow. There is no paid bridge. Migration is the only exit.

4. Rotate any credentials an exposed server could have leaked. If a tracking server below 3.15.0 was internet-reachable, treat its instance credentials as compromised — the exploit reads cloud metadata.

Track MLflow — and everything else in your stack

MLflow, Ray and the rest of your stack’s end-of-life status, risk scores and KEV exposure — checked daily.

MLflow lifecycle & risk →

Frequently asked questions

Is MLflow 2.x end of life?
MLflow has never published a lifecycle policy or an end-of-life announcement for the 2.x line. In practice it is end of life: the final 2.x release, 2.22.5, shipped 2026-05-12, and the critical actively-exploited SSRF CVE-2026-64849 was fixed only in 3.15.0 with no 2.x backport. An unpatched, actively exploited line that receives no fixes is end of life whether or not anyone announces it.
Is MLflow affected by CVE-2026-64849?
Every MLflow release below 3.15.0 is vulnerable, including all of 2.x and 1.x. GitHub advisory GHSA-7gwp-5pfp-969j records the vulnerable range as below 3.15.0 and the first patched version as 3.15.0. CISA added the flaw to the Known Exploited Vulnerabilities catalog on August 19, 2026 with a federal remediation deadline of September 2, 2026.
What version of MLflow fixes CVE-2026-64849?
MLflow 3.15.0 is the first patched version, per GitHub advisory GHSA-7gwp-5pfp-969j. MLflow's own security advisory does not name a patched version, so the GitHub Security Advisory database is the authoritative record here. Upgrade to 3.15.0 or later; there is no fixed 2.x release.
Is there extended support for MLflow 2.x?
No. None of the commercial extended-lifecycle-support vendors we track offer post-EOL patching for MLflow. Unlike PHP or Java, where a paid bridge exists, the only fix for an MLflow 2.x deployment is migrating to 3.15.0 or later.
Why are AI tools suddenly appearing in CISA's KEV catalog?
MLflow is the third AI infrastructure tool added to CISA's Known Exploited Vulnerabilities catalog in six weeks: Langflow entries continued through August 4, 2026 (six KEV entries since May 2025), Ray's CVE-2025-62593 was added August 17, 2026, and MLflow's CVE-2026-64849 followed on August 19, 2026. ML tooling typically runs with cloud credentials and rarely sits behind the same hardening as production web apps, which makes it a productive target.