Description
A security vulnerability has been detected in DLR-RM stable-baselines3 up to 2.9.0. This affects the function PPO.load/load_replay_buffer/VecNormalize.load of the file save_util.py. Such manipulation leads to deserialization. It is possible to launch the attack remotely. The exploit has been disclosed publicly and may be used. In v2.9.0 the PyTorch tensor load path is hardened (weights_only=True), but that hardening was later reverted on master via PR #1913 "Hotfix: revert loading with weights_only=True" [blocked] to fix PyTorch 1.13 compat - so even the one "safe" path is inconsistent across versions. #2281 was closed as a duplicate of #1831 since both are unsafe pickle deserialization - but #1831's fix (PR #41) only gated the Hugging Face Hub loader in the separate huggingface_sb3 package. This finding covers the core stable_baselines3 load APIs (PPO.load, load_replay_buffer, VecNormalize.load), which have no safe mode or gate and remained exploitable in v2.9.0 until the outstanding hardening (PR #2264) ships.
Exploitation Probability (EPSS)
Low PriorityThe Exploit Prediction Scoring System (EPSS) uses machine learning to estimate the probability that a vulnerability will be exploited in the wild within the next 30 days.
7-Day Exploitation Trend
Vulnerability Timeline
1 eventsThreat Actor Attribution
PREMIUM INTELRemediation & Mitigation
SOLUTIONOfficial patches and mitigation steps are available for this vulnerability.
apt-get update && apt-get upgrade -y specific-package
# Verify installation
dpkg -l | grep package-name
Am I Vulnerable?
Check your domain or package.json for CVE-2026-94093 exposure.
Vulnerability Details
Need Manual Validation?
Automated scanners flag false positives. Get a manual pentest validation for this CVE.