GATEBLEED: Exploiting On-Core Accelerator Power Gating for High Performance and Stealthy Attacks on AI

As power consumption from AI training and inference continues to increase, AI accelerators are being integrated directly into the CPU. Intel's Advanced Matrix Extensions AMX is one such example, debuting on the 4th generation Intel Xeon Scalable CPU. We discover a timing side and covert channel,...

Optimizing Canaries for Privacy Auditing with Metagradient Descent

In this work we study black-box privacy auditing, where the goal is to lower bound the privacy parameter of a differentially private learning algorithm using only the algorithm's outputs i.e., final trained model. For DP-SGD the most successful method for training differentially private deep...

Amazon warns 200 million Prime customers that scammers are after their login info

Amazon has sent out an alert to its 200 million customers, warning them that scammers are impersonating Amazon in a Prime membership scam. In the email, sent earlier this month, Amazon said it had noticed an increase in reports about fake Amazon emails: What 's happening: Scammers are sending fak...

Vulnerabilities fixed in Splunk Enterprise and Splunk Cloud Platform

Splunk has fixed vulnerabilities in Splunk Enterprise and Splunk Cloud Platform. The vulnerabilities in Splunk Enterprise and Splunk Cloud Platform allow both low-privileged and high-privileged users to perform unauthorized actions, such as suppressing alerts, executing remote commands, and causi...

CVE-2025-20321

In Splunk Enterprise versions below 9.4.3, 9.3.5, 9.2.7 and 9.1.10, and Splunk Cloud Platform versions below 9.3.2411.104, 9.3.2408.114, and 9.2.2406.119, an unauthenticated attacker can send a specially-crafted SPL search that could change the membership state in a Splunk Search Head Cluster SHC...

CVE-2025-20321 Membership State Change in Splunk Search Head Cluster through a Cross-Site Request Forgery (CSRF) in Splunk Enterprise

In Splunk Enterprise versions below 9.4.3, 9.3.5, 9.2.7 and 9.1.10, and Splunk Cloud Platform versions below 9.3.2411.104, 9.3.2408.114, and 9.2.2406.119, an unauthenticated attacker can send a specially-crafted SPL search that could change the membership state in a Splunk Search Head Cluster SHC...

CVE-2025-20321 Membership State Change in Splunk Search Head Cluster through a Cross-Site Request Forgery (CSRF) in Splunk Enterprise

In Splunk Enterprise versions below 9.4.3, 9.3.5, 9.2.7 and 9.1.10, and Splunk Cloud Platform versions below 9.3.2411.104, 9.3.2408.114, and 9.2.2406.119, an unauthenticated attacker can send a specially-crafted SPL search that could change the membership state in a Splunk Search Head Cluster SHC...

Security update for samba

This update for samba fixes the following issues: CVE-2025-0620: smbd doesn't pick up group membership changes when re-authenticating an expired SMB session bsc1244136. Other bugfixes: net ad join fails with "Failed to join domain: failed to create kerberos keytab" bsc1238063. Patch Instructions:...

SUSE-SU-2025:02230-1 Security update for samba

This update for samba fixes the following issues: - CVE-2025-0620: smbd doesn't pick up group membership changes when re-authenticating an expired SMB session bsc1244136. Other bugfixes: - net ad join fails with 'Failed to join domain: failed to create kerberos keytab' bsc1238063...

CVE-2025-47871

Mattermost versions 10.5.x = 10.5.5, 9.11.x = 9.11.15, 10.8.x = 10.8.0, 10.7.x = 10.7.2, 10.6.x = 10.6.5 fail to properly validate channel membership when retrieving playbook run metadata, allowing authenticated users who are playbook members but not channel members to access sensitive informatio...

GHSA-WGVP-JJ4W-88HF Mattermost Incorrect Authorization vulnerability

Mattermost versions 10.5.x = 10.5.5, 9.11.x = 9.11.15, 10.8.x = 10.8.0, 10.7.x = 10.7.2, 10.6.x = 10.6.5 fail to properly validate channel membership when retrieving playbook run metadata, allowing authenticated users who are playbook members but not channel members to access sensitive informatio...

Mattermost Incorrect Authorization vulnerability

Mattermost versions 10.5.x = 10.5.5, 9.11.x = 9.11.15, 10.8.x = 10.8.0, 10.7.x = 10.7.2, 10.6.x = 10.6.5 fail to properly validate channel membership when retrieving playbook run metadata, allowing authenticated users who are playbook members but not channel members to access sensitive informatio...

CVE-2025-47871

Mattermost versions 10.5.x = 10.5.5, 9.11.x = 9.11.15, 10.8.x = 10.8.0, 10.7.x = 10.7.2, 10.6.x = 10.6.5 fail to properly validate channel membership when retrieving playbook run metadata, allowing authenticated users who are playbook members but not channel members to access sensitive informatio...

CVE-2025-47871

Mattermost requires update: versions 10.5.x ≤ 10.5.5, 9.11.x ≤ 9.11.15, 10.8.x ≤ 10.8.0, 10.7.x ≤ 10.7.2, 10.6.x ≤ 10.6.5 fail to validate channel membership when retrieving playbook run metadata. This allows authenticated users who are playbook members but not channel members to access sensitive...

CVE-2025-47871 Mattermost Playbooks exposes private channel metadata to unauthorized users via run metadata API

Mattermost versions 10.5.x = 10.5.5, 9.11.x = 9.11.15, 10.8.x = 10.8.0, 10.7.x = 10.7.2, 10.6.x = 10.6.5 fail to properly validate channel membership when retrieving playbook run metadata, allowing authenticated users who are playbook members but not channel members to access sensitive informatio...

Amplifying Machine Learning Attacks through Strategic Compositions

Machine learning ML models are proving to be vulnerable to a variety of attacks that allow the adversary to learn sensitive information, cause mispredictions, and more. While these attacks have been extensively studied, current research predominantly focuses on analyzing each attack type...

Image Corruption-Inspired Membership Inference Attacks against Large Vision-Language Models

Large vision-language models LVLMs have demonstrated outstanding performance in many downstream tasks. However, LVLMs are trained on large-scale datasets, which can pose privacy risks if training images contain sensitive information. Therefore, it is important to detect whether an image is used t...

Enhancing One-run Privacy Auditing with Quantile Regression-Based Membership Inference

Differential privacy DP auditing aims to provide empirical lower bounds on the privacy guarantees of DP mechanisms like DP-SGD. While some existing techniques require many training runs that are prohibitively costly, recent work introduces one-run auditing approaches that effectively audit DP-SGD...

SOFT: Selective Data Obfuscation for Protecting LLM Fine-Tuning against Membership Inference Attacks

Whitepaper called SOFT: Selective Data Obfuscation For Protecting LLM Fine-Tuning Against Membership Inference Attacks...

CVE-2025-49265

Missing Authorization vulnerability in WP Swings Membership For WooCommerce membership-for-woocommerce allows Accessing Functionality Not Properly Constrained by ACLs.This issue affects Membership For WooCommerce: from n/a through = 2.8.1...