NewApp Monitoring is liveSign up free
AgentbeamBeta
Open-source package vulnerability · PyPI

CVE-2026-114608: Pylint: Arbitrary code execution via pickle deserialization of the stats cache in PYLINT_HOME

mediumCVSS 4.0 5.4GHSA-22qq-2fwx-7p67

Pylint before 4.0.9 saved its stats cache as pickle files in `PYLINT_HOME` and loaded them with `pickle.load` on the next run. Someone who can write to that directory can plant a crafted pickle and run code as the user running pylint. Upgrade to 4.0.9, which stores the cache as JSON.

Published Oct 27, 2026 · Analysis Nov 1, 2026

What an attacker can do

Since pylint 0.15.2, in 2006, the linter has written run statistics as pickles under PYLINT_HOME and read them back at the start of later runs. Unpickling data can run arbitrary code. An attacker who cannot touch your code or configuration, but can write into the cache directory, drops a malicious pickle there. The next pylint run loads it and the payload executes with that user's rights.

The realistic case is a multi-user host or shared CI runner where PYLINT_HOME is predictable, shared, or group-writable. The default, ~/.cache/pylint, is writable only by the user running pylint, so a standard single-user setup is not exposed. The CVSS 4.0 vector reflects local access, high complexity and a required victim action. Impact on the lint host is high for confidentiality, integrity and availability, so CI secrets are the usual prize.

Who is affected

The pylint pip package before 4.0.9. You are at risk when PYLINT_HOME points at a directory that other users, jobs or untrusted processes can write to.

Check the version and the cache location:

pip show pylint
echo $PYLINT_HOME

Then check the directory's permissions, for example with ls -ld on that path.

pylint

Ecosystem
PyPI
Affected
< 4.0.9
Fixed in
4.0.9
Check
pip show pylint

How to fix it

Upgrade to pylint 4.0.9.

pip install -U "pylint>=4.0.9"

From 4.0.9 the result cache is stored as JSON instead of pickle, which removes the code execution path. Until you upgrade, do not point PYLINT_HOME to an untrusted, shared or group-writable directory. After upgrading, delete old pickle files from the cache directory.

How to tell if you were hit

Inspect PYLINT_HOME (default ~/.cache/pylint) for pickle files you did not expect, with owners other than the pylint user or unusual modification times. On shared runners, look for writes to that directory by other accounts, and for pylint processes that spawn unexpected child processes or network connections.

Original advisory

From the GitHub Advisory Database (CC BY 4.0).

Impact

Since 2006 / pylint 0.15.2, pylint caches its stats as pickles in PYLINT_HOME and reload them on the next run with pickle.load. Someone without access to the configuration or linted code, but with access to the cache directory (predictable PYLINT_HOME on a multi-user host) can write a crafted pickle that will runs arbitrary code when pylint access its stat cache.

Patches

From pylint 4.0.9, the result cache is now stored as JSON instead of pickle, preventing code-execution.

Workarounds

Until upgrading do not point PYLINT_HOME to an untrusted, shared, or group-writable directory. The default value, ~/.cache/pylint, is writable only by the user running pylint.

Credits

  • Reported privately via Tidelift coordinated disclosure (Dan Lenz).

Questions

Is the default pylint setup vulnerable to CVE-2026-114608?

Only if someone else can write to the cache. The advisory says the default ~/.cache/pylint is writable only by the user running pylint.

Which pylint version fixes it?

4.0.9. The cache is now JSON, not pickle.

Does the attacker need to change my code?

No. They only need write access to the cache directory. That is why shared runners and group-writable cache paths are the risk.

References

How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. The analysis is written by an AI model from the sources listed above and checked by automated rules before it is published. Check versions against the vendor's advisory before you act. Methodology