CVE-2026-94093Medium· 6.3▾ TwilightPoC availableA 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 possi…
▾ Twilight zone — High severity, or a signal on a lesser flaw
impact 34.7 · likelihood 0.1 · exploitation 12
A public proof-of-concept already exists for this vulnerability — see Exploit availability below.
Stakeholder-Specific Vulnerability Categorization from CISA's ADP record at CVE.org: whether exploitation is observed, whether an attack can be automated, and how much of the system is at stake.
Disclosure to exploitation, from the record and what we observed since indexing it.
Disclosed via NVD
Last analysed / modified upstream
0.3%
Exploit / PoC code exists
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.
Refer to the linked advisories for vendor-supplied fixes and affected version ranges.
Field changes observed since this record was first indexed.
Connected by shared product, vendor, weakness, or advisory.
CVE-2021-45105Medium· 5.9Apache Log4j2 versions 2.0-alpha1 through 2.16.0 (excluding 2.12.3 and 2.3.1) did not protect from uncontrolled recursion from self-referential lookups
CVE-2015-2291High· 7.8(1) IQVW32.sys before 1.3.1.0 and (2) IQVW64.sys before 1.3.1.0 in the Intel Ethernet diagnostics driver for Windows allows local users to cause a denial of service or possibly execute arbitrary code with kernel privileges via a crafted …
CVE-2025-6558High· 8.8Insufficient validation of untrusted input in ANGLE and GPU in Google Chrome prior to 138.0.7204.157 allowed a remote attacker to potentially perform a sandbox escape via a crafted HTML page
CVE-2025-34115High· 8.7An authenticated command injection vulnerability exists in OP5 Monitor through version 7.1.9 via the 'cmd_str' parameter in the command_test.php endpoint
CVE-2025-13805Low· 3.7A weakness has been identified in nutzam NutzBoot up to 2.6.0-SNAPSHOT
CVE-2020-3478High· 8.1A vulnerability in the REST API of Cisco Enterprise NFV Infrastructure Software (NFVIS) could allow an authenticated, remote attacker to overwrite certain files that should be restricted on an affected device