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.
Published: 2026-09-20
Score: 5.3 Medium
EPSS: < 1% Very Low
KEV: No
Impact: n/a
Action: n/a
AI Analysis

Analysis and contextual insights are available on OpenCVE Cloud.

Remediation

No vendor fix or workaround currently provided.

Additional remediation guidance may be available on OpenCVE Cloud.

Tracking

Sign in to view the affected projects.

Advisories

No advisories yet.

History

Tue, 22 Sep 2026 02:30:00 +0000

Type Values Removed Values Added
Metrics ssvc

{'options': {'Automatable': 'no', 'Exploitation': 'poc', 'Technical Impact': 'partial'}, 'version': '2.0.3'}

Sun, 20 Sep 2026 23:00:00 +0000

Type Values Removed Values Added
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.
Title DLR-RM stable-baselines3 save_util.py VecNormalize.load deserialization
First Time appeared Dlr-rm
Dlr-rm stable-baselines3
Weaknesses CWE-20
CWE-502
CPEs cpe:2.3:a:dlr-rm:stable-baselines3:*:*:*:*:*:*:*:*
Vendors & Products Dlr-rm
Dlr-rm stable-baselines3
References
Metrics cvssV2_0

{'score': 7.5, 'vector': 'AV:N/AC:L/Au:N/C:P/I:P/A:P/E:POC/RL:ND/RC:UR'}

cvssV3_0

{'score': 6.3, 'vector': 'CVSS:3.0/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R'}

cvssV3_1

{'score': 6.3, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R'}

cvssV4_0

{'score': 5.3, 'vector': 'CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N/E:P'}

Subscriptions

Dlr-rm Stable-baselines3
cve-icon MITRE

Status: PUBLISHED

Assigner: VulDB

Published:

Updated: 2026-09-21T16:21:48.911Z

Reserved: 2026-09-20T08:39:03.152Z

Link: CVE-2026-94093

cve-icon Vulnrichment

Updated: 2026-09-21T16:21:29.305Z

cve-icon NVD

Status : Deferred

Published: 2026-09-20T23:17:03.343

Modified: 2026-09-21T17:19:19.420

Link: CVE-2026-94093

cve-icon Redhat

No data.

cve-icon OpenCVE Enrichment

Updated: 2026-09-21T10:01:58Z

Weaknesses