7183 matches found
Attacking Attention of Foundation Models Disrupts Downstream Tasks
Foundation models represent the most prominent and recent paradigm shift in artificial intelligence. Foundation models are large models, trained on broad data that deliver high accuracy in many downstream tasks, often without fine-tuning. For this reason, models such as CLIP , DINO or Vision...
SECNEURON: Reliable and Flexible Abuse Control in Local LLMs Via Hybrid Neuron Encryption
Large language models LLMs with diverse capabilities are increasingly being deployed in local environments, presenting significant security and controllability challenges. These locally deployed LLMs operate outside the direct control of developers, rendering them more susceptible to abuse...
COALESCE: Economic and Security Dynamics of Skill-Based Task Outsourcing among Team of Autonomous LLM Agents
The meteoric rise and proliferation of autonomous Large Language Model LLM agents promise significant capabilities across various domains. However, their deployment is increasingly constrained by substantial computational demands, specifically for Graphics Processing Unit GPU resources. This pape...
ASB-A-383080440
In loadDrawableForCookie of ResourcesImpl.java, there is a possible way to access task snapshots of other apps due to a confused deputy. This could lead to local escalation of privilege with no additional execution privileges needed. User interaction is not needed for exploitation...
Computer Vision Annotation Tool 安全漏洞
Computer Vision Annotation Tool CVAT is a cvat.ai open source interactive video and image annotation tool for computer vision. A security vulnerability exists in Computer Vision Annotation Tool versions prior to 2.4.0 through 2.38.0, which stems from the possibility that an authenticated user may...
Safety Alignment Can Be Not Superficial with Explicit Safety Signals
Recent studies on the safety alignment of large language models LLMs have revealed that existing approaches often operate superficially, leaving models vulnerable to various adversarial attacks. Despite their significance, these studies generally fail to offer actionable solutions beyond data...
DaaS - Change master image it fails with "ProvisioningTaskError"
Unable to update DaaS Machine Catalog - Access Machine Catalog "Change master image" it fails with "ProvisioningTaskError" ErrorMessage - HandleExplicitStorage Failed Error retrieving item from path ""...
Linux kernel 安全漏洞
Linux kernel is the kernel used by Linux, the open source operating system of the Linux Foundation in the United States. A security vulnerability exists in the Linux kernel that stems from the UCSI driver not waiting for all tasks to complete before partner removal...
kernel: sched/rt: Fix race in push_rt_task
In the Linux kernel, the following vulnerability has been resolved: sched/rt: Fix race in pushrttask Overview ======== When a CPU chooses to call pushrttask and picks a task to push to another CPU's runqueue then it will call findlocklowestrq method which would take a double lock on both CPUs'...
kernel: sched/rt: Fix race in push_rt_task
In the Linux kernel, the following vulnerability has been resolved: sched/rt: Fix race in pushrttask Overview ======== When a CPU chooses to call pushrttask and picks a task to push to another CPU's runqueue then it will call findlocklowestrq method which would take a double lock on both CPUs'...
Updating MCS catalog fails with error FailedToUnregisterMasterImageFromMdm
Error updating an MCS catalog in Citrix DAAS Action Name: MCCreateMachineCatalogInitialzation Exception: StudioErrorId : ProvisioningTaskError ErrorCategory : NotSpecified TaskState : FailedToUnregisterMasterImageFromMdm TaskStateInformation : Terminated ErrorId :...
CVE-2025-1415
A low-privileged user is able to obtain information about tasks executed on devices controlled by Proget MDM Mobile Device Management, as well as details of the devices like their UUIDs needed for exploitation of CVE-2025-1416. In order to perform the attack, one has to know a taskid, but since...
CVE-2025-0218
When batch jobs are executed by pgAgent, a script is created in a temporary directory and then executed. In versions of pgAgent prior to 4.2.3, an insufficiently seeded random number generator is used when generating the directory name, leading to the possibility for a local attacker to pre-creat...
CVE-2025-0464
A vulnerability was found in SourceCodester Task Reminder System 1.0. It has been declared as problematic. Affected by this vulnerability is an unknown functionality of the component Maintenance Section. The manipulation of the argument System Name leads to cross site scripting. The attack can be...
CVE-2024-48706
Collabtive 3.1 is vulnerable to Cross-site scripting XSS via the title parameter with action=add or action=editform within the a managemessage.php file and b managetask.php file respectively...
CVE-2024-46081
Scriptcase v9.10.023 and before is vulnerable to Cross Site Scripting XSS. An authenticated user can craft malicious payloads in the To-Do List. The assigned user will trigger a stored XSS, which is particularly dangerous because tasks are assigned to various users on the platform...
CVE-2024-8140
A vulnerability was found in SourceCodester Task Progress Tracker 1.0 and classified as problematic. Affected by this issue is some unknown functionality of the file update-task.php. The manipulation of the argument taskname leads to cross site scripting. The attack may be launched remotely. The...
CVE-2024-7792
A vulnerability was found in SourceCodester Task Progress Tracker 1.0. It has been classified as critical. Affected is an unknown function of the file /endpoint/delete-task.php. The manipulation of the argument task leads to sql injection. It is possible to launch the attack remotely. The exploit...
CVE-2024-7793
A vulnerability was found in SourceCodester Task Progress Tracker 1.0. It has been declared as problematic. Affected by this vulnerability is an unknown functionality of the file /endpoint/add-task.php. The manipulation of the argument taskname leads to cross site scripting. The attack can be...