Code execution via unrestricted deserialization imports
Published Mar 11, 2025 · Updated Jul 24, 2025
Code injection in Google Keras 3.0.0 through 3.7.0 allows attackers to execute Python code when a crafted .keras model is loaded. The _retrieve_class_or_fn helper imports any module named by config.json and returns its named function, while functional model reconstruction invokes that function with attacker-controlled inbound-node arguments. Loading an untrusted model is required; execution occurs during parsing with the Keras process permissions even when safe_mode is enabled.
Summary
What happened
Code injection in Google Keras 3.0.0 through 3.7.0 allows attackers to execute Python code when a crafted .keras model is loaded. The _retrieve_class_or_fn helper imports any module named by config.json and returns its named function, while functional model reconstruction invokes that function with attacker-controlled inbound-node arguments. Loading an untrusted model is required; execution occurs during parsing with the Keras process permissions even when safe_mode is enabled.
The record
- CVE
- CVE-2025-1550
- Published
- Mar 11, 2025
- Updated
- Jul 24, 2025
- Vendor
- Product
- Keras
- Classifications
- CWE-94, CAPEC-88, T1204.002
- Attack vector
- local
- Privileges
- Unavailable
Timeline
How it unfolded
- Mar 11, 2025CVE publishedPublication date reported by the CVE source.
- Jul 24, 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 <3.8.0
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
- CVE-2025-1550 crafted .keras model proof of conceptproof of concept · 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