Browse / Data and Model Poisoning / CVE-2026-10803
MLflow: Deterministic sampling in dataset digest enables predictable collisions
- Class
- Data and Model Poisoning (filed by a model that read the advisory)
- Technique
- Other weaknesses
- Also involves
- Unsafe Model Loading and Supply Chain
- Affected
- mlflow (PyPI)
- CVSS
- CVSS:3.1/AV:L/AC:H/PR:L/UI:N/S:U/C:N/I:L/A:L
- EPSS
- 0.1% chance of exploitation in 30 days
How it works, in short
Dataset digests use deterministic sampling with a weak hash, so an attacker can craft a different dataset that produces the same digest. (Our summary, written by a model. The source text is below.)
Details
A flaw has been found in MLflow up to 3.10.0. This issue affects the function mlflow.data.digest_utils of the file mlflow/data/digest_utils.py of the component Dataset Digest Computation. This manipulation causes use of weak hash. It is possible to launch the attack on the local host. The attack is considered to have high complexity. The exploitability is assessed as difficult. The exploit has been published and may be used. The project was informed of the problem early through a pull request but has not reacted yet.
References
- https://nvd.nist.gov/vuln/detail/CVE-2026-10803
- https://github.com/mlflow/mlflow/issues/22419
- https://github.com/mlflow/mlflow/pull/22420
- https://github.com/mlflow/mlflow
- https://github.com/pypa/advisory-database/tree/main/vulns/mlflow/PYSEC-2026-195.yaml
- https://vuldb.com/cve/CVE-2026-10803
- https://vuldb.com/submit/831462
- https://vuldb.com/vuln/368252
- https://vuldb.com/vuln/368252/cti
Source: GitHub Advisory Database, used under CC BY 4.0.