AIKIDO-2026-319121
pyod is vulnerable to Deserialization of Untrusted Data
63
Medium Risk
This Affects:
pyodTL;DR
The pyod.utils.persistence module exposes load and compat_load helpers that restore saved detector models by deserializing them with joblib. Before the fix, both helpers unpickled the artifact immediately, so the envelope, schema, and strict-mode checks ran only after the pickle stream had already executed its embedded reducers. Loading a crafted model file therefore runs arbitrary code in the host process. The fix requires callers to pass an explicit trusted acknowledgement before either helper deserializes an artifact.
Who does this affect?
You are affected if you are using a version that falls within the vulnerable range and your application loads pyod model artifacts from untrusted or unverified sources.
Background info
pyod is vulnerable to Deserialization of Untrusted Data in versions 3.5.0 - 3.6.1.
How to fix this
Upgrade the pyod library to the patch version.
Links
GitHub Release
Other
github.com/yzhao062/pyod/
pypi.org/project/pyod/3.6.2/vuldb.com/cve/CVE-2026-15529vuldb.com/submit/854559vuldb.com/vuln/377872vuldb.com/vuln/377872/cti
Are You Affected?
Connect your repositories to instantly see whether vulnerable or malicious packages exist in your codebase.
Free. No credit card required.
I consent to receiving marketing communications based on Aikido’s Privacy Policy.
SOC 2Compliant
ISO 27001CompliantISO 42001Compliant
FedRAMPImplementing© All Intel data is openly available and commercially licensed.