1487 matches found
heist-salt (>=1.0.0 <=5.3.1), isalt (>=2019.12.0 <=2021.2.2) +9 more potentially affected by CVE-2024-38823 via salt (=3007.14.0)
salt PYPI version =3007.14.0 is affected by a known vulnerability. The following packages have a transitive dependency on salt and may be impacted: - heist-salt =1.0.0, =2019.12.0, =2019.6.28a3, =2.6.2.dev0, =2020.7.0, =1.1.1, =22.9.8.0rc1, =0.3.2, =0.5.7 - slskit =2022.4.0 Source cves:...
Learning Obfuscations of LLM Embedding Sequences: Stained Glass Transform
The high cost of ownership of AI compute infrastructure and challenges of robust serving of large language models LLMs has led to a surge in managed Model-as-a-service deployments. Even when enterprises choose on-premises deployments, the compute infrastructure is typically shared across many tea...
nautobot-chatops (>=1.6.0 <=1.7.1), nautobot-chatops-arista-cloudvision (>=1.0.1 <=1.3.0) +7 more potentially affected by CVE-2025-49143 via nautobot (>=1.0.3 <=1.5.16)
nautobot PYPI version =1.0.3, =1.6.0, =1.0.1, =1.1.0, =0.9.2, =1.5.0, =0.9.0, =0.1.0, =0.1.0, =0.2.0 Source cves: CVE-2025-49143 Source advisory: OSV:GHSA-RH67-4C8J-HJJH...
LLM Unlearning Should Be Form-Independent
Large Language Model LLM unlearning aims to erase or suppress undesirable knowledge within the model, offering promise for controlling harmful or private information to prevent misuse. However, recent studies highlight its limited efficacy in real-world scenarios, hindering practical adoption. In...
3D Gaussian Splat Vulnerabilities
With 3D Gaussian Splatting 3DGS being increasingly used in safety-critical applications, how can an adversary manipulate the scene to cause harm? We introduce CLOAK, the first attack that leverages view-dependent Gaussian appearances - colors and textures that change with viewing angle - to embed...
alaas (>=0.1.6 <=0.2.1), annlite (>=0.3.14 <=0.4.0) +73 more potentially affected by CVE-2025-5150 via docarray (>=0.12.9 <=0.41.0)
docarray PYPI version =0.12.9, =0.1.6, =0.3.14, =0.0.3, =0.1.0, =0.1.0, =0.1.7, =0.3.9, =0.3.9, =0.5.2, =0.3.9, =0.3.9, =0.5.2, =0.5.7 - fasr-enhancement-deepfilternet =0.5.7 - fasr-enhancement-dtln =0.5.7 and more Source cves: CVE-2025-5150 Source advisory: SNYK:PYTHON-DOCARRAY-10246594...
CVE-2023-36627
A flaw exists in FlashBlade Purity whereby a user with access to an administrative account on a FlashBlade that is configured with timezone-dependent snapshot schedules can configure a timezone to prevent the schedule from functioning properly...
CVE-2021-46038
A Pointer Dereference vulnerability exists in GPAC 1.0.1 in unlinkchunk.isra, which causes a Denial of Service context-dependent...
CVE-2021-46044
A Pointer Dereference Vulnerabilty exists in GPAC 1.0.1via ShiftMetaOffset.isra, which causes a Denial of Service context-dependent...
CVE-2012-4028
Tridium Niagara AX Framework does not properly store credential data, which allows context-dependent attackers to bypass intended access restrictions by using the stored information for authentication...
CVE-2011-5126
Blue Coat ProxySG 6.1 before SGOS 6.1.5.1 and 6.2 before SGOS 6.2.2.1 writes the secure heap to core images, which allows context-dependent attackers to obtain sensitive authentication information by leveraging read access to a downloaded core file...
CVE-2011-4723
The D-Link DIR-300 router stores cleartext passwords, which allows context-dependent attackers to obtain sensitive information via unspecified vectors...
CVE-2014-2870
The default configuration of PaperThin CommonSpot before 7.0.2 and 8.x before 8.0.3 uses cleartext for storage of credentials in a database, which makes it easier for context-dependent attackers to obtain sensitive information via unspecified vectors...
CVE-2005-4895
Multiple integer overflows in TCMalloc tcmalloc.cc in gperftools before 0.4 make it easier for context-dependent attackers to perform memory-related attacks such as buffer overflows via a large size value, which causes less memory to be allocated than expected...
CVE-2005-4849
Apache Derby before 10.1.2.1 exposes the 1 user and 2 password attributes in cleartext via a the RDBNAM parameter of the ACCSEC command and b the output of the DatabaseMetaData.getURL function, which allows context-dependent attackers to obtain sensitive information...
SUSE CVE-2025-4658
Versions of OpenPubkey library prior to 0.10.0 contained a vulnerability that would allow a specially crafted JWS to bypass signature verification. As OPKSSH depends on the OpenPubkey library for authentication, this vulnerability in OpenPubkey also applies to OPKSSH versions prior to 0.5.0 and...
CVE-2025-4638
A vulnerability exists in the inftrees.c component of the zlib library, which is bundled within the PointCloudLibrary PCL. This issue may allow context-dependent attackers to cause undefined behavior by exploiting improper pointer arithmetic. Since version 1.14.0, PCL by default uses a zlib...
CVE-2025-4638 Improper Pointer Arithmetic in pcl
A vulnerability exists in the inftrees.c component of the zlib library, which is bundled within the PointCloudLibrary PCL. This issue may allow context-dependent attackers to cause undefined behavior by exploiting improper pointer arithmetic. Since version 1.14.0, PCL by default uses a zlib...
MTL-UE: Learning to Learn Nothing for Multi-Task Learning
Most existing unlearnable strategies focus on preventing unauthorized users from training single-task learning STL models with personal data. Nevertheless, the paradigm has recently shifted towards multi-task data and multi-task learning MTL, targeting generalist and foundation models that can...
