Code execution via ignored HDF5 safe mode
Published Sep 19, 2025 · Updated Sep 19, 2025
Code execution in Keras 3.0.0 through 3.11.2 allows attackers to run Python code by inducing a user to load a crafted HDF5 model. The load_model path delegates .h5 and .hdf5 inputs to legacy_h5_format.load_model_from_hdf5 without passing or checking safe_mode, permitting Lambda-layer pickled code to deserialize. Exploitation requires a user or application to load an attacker-supplied legacy model; the payload runs with the Keras process's privileges.
Summary
What happened
Code execution in Keras 3.0.0 through 3.11.2 allows attackers to run Python code by inducing a user to load a crafted HDF5 model. The load_model path delegates .h5 and .hdf5 inputs to legacy_h5_format.load_model_from_hdf5 without passing or checking safe_mode, permitting Lambda-layer pickled code to deserialize. Exploitation requires a user or application to load an attacker-supplied legacy model; the payload runs with the Keras process's privileges.
The record
- CVE
- CVE-2025-9905
- Published
- Sep 19, 2025
- Updated
- Sep 19, 2025
- Vendor
- Keras
- Product
- Keras
- Classifications
- CWE-913, CAPEC-175, T1204.002
- Attack vector
- local
- Privileges
- unauthenticated
Timeline
How it unfolded
- Sep 19, 2025CVE publishedPublication date reported by the CVE source.
- Sep 19, 2025Record updatedLatest update available in the CVE record.
Exploitability
Present is not the same as exploitable
Compare your product and version with the public record. A matching version still requires validation against your environment.
Is a vulnerable build present?
Compare these published version ranges with your installed build and any vendor patches.
- Affected versionversion=3.0.0 <=3.11.2
What conditions does exploitation require?
What is affected?
Attacks
What attackers are doing with it
Daily unique IPs observed by Shadowserver honeypots for known exploited vulnerabilities (KEVs). Missing observations do not establish an absence of attacks.
Weakness, pattern, technique
Public exploit references
- GHSA-36rr-ww3j-vrjv HDF5 model proof of conceptfunctional · demonstrated
Labels summarize the accepted research assessment. They do not indicate a test against your environment.
Technologies
Your stack
See the directory against your own environment.
Your stack
Check the software in your environment
Book a demo to see how Hinoki identifies affected software and validates exploitability in your environment.
Book a demo