Falcon Secure Access Sets the Standard for Zero Trust Browser Security
Falcon Secure Access Sets the Standard for Zero Trust Browser Security
2026-08-09 归档
今日共收录 815 条安全情报，包含 75 个 CVE，398 篇安全通告，以及 81 篇研究论文。
← 返回归档列表Beyond the Model: Harnessing Frontier AI for Stronger Cyber Defense
Beyond the Model: Harnessing Frontier AI for Stronger Cyber Defense
make smarter security investments and mitigate future risks
make smarter security investments and mitigate future risks
Inside Astaroth's New Spambot Component
Inside Astaroth's New Spambot Component
Falcon Platform IOAs Arrive in Falcon Next-Gen SIEM to Identify New Threats
Falcon Platform IOAs Arrive in Falcon Next-Gen SIEM to Identify New Threats
CrowdStrike 2026 Threat Hunting Report: Exploitation Window Closes as AI Use Accelerates
CrowdStrike 2026 Threat Hunting Report: Exploitation Window Closes as AI Use Accelerates
Secure Agent Harness Execution: Preventing EscapeAug 04, 2026
Secure Agent Harness Execution: Preventing EscapeAug 04, 2026
Expanding AI Benchmarks in Cybersecurity Beyond Vulnerability DiscoveryAug 06, 2026
Expanding AI Benchmarks in Cybersecurity Beyond Vulnerability DiscoveryAug 06, 2026
CrowdStrike Threat Hunts for Shell Command Obfuscation on VMware ESX
CrowdStrike Threat Hunts for Shell Command Obfuscation on VMware ESX
Metabase SQLi zero-day exploited in customer data-theft attacks
A critical Metabase SQL injection vulnerability was exploited in zero-day attacks to breach customer instances in data t
Unlimited Technology Systems breach impacts 3.8 million people
Healthcare software company Unlimited Technology Systems reported that more than 3.8 million people were impacted by a d
Levi Strauss & Co. says hackers stole corporate data in cyberattack
Levi Strauss & Co. (Levi's) says that hackers used social engineering on three of its employees to gain access to and st
ClickFix Attacks Deliver macOS Stealer That Can Drain Crypto Wallets
ClickFix-style attacks are being used to deliver a Go-based malware capable of stealing cryptocurrency assets, as well a
UNC6671 Vishing Attacks Target Personal Phones to Steal SaaS Data
A recent wave of cyber attacks targeting financial services, private equity, and professional services has been attribut
Hackers breach TrueConf to trojanize client installers with backdoors
The Head Mare hacktivist group has been exploiting vulnerabilities in unpatched TrueConf video conferencing servers to r
New CSS Attacks Can Break Webmail Defenses to Steal Passwords and Tokens
New research shows content inside an email can escape its message boundary and interfere with the webmail interface. Ac
When Experience Becomes Instruction: Trajectory Poisoning in Self-Evolving Agent Skill Systems
本文研究了自进化技能（Self-Evolving Skill, SES）系统中的轨迹投毒攻击。SES 系统将智能体（agent）的执行轨迹蒸馏为持久化技能，使得原本不可信的经验能够被提升为可信的指令。这种机制虽然提升了智能体的适应能力，但也
Breaking Customized LLMs for Coding: Automated Red Teaming for Instruction Backdoor Attacks
本文针对大模型定制化平台（如通过系统提示词嵌入指令来构建代码智能任务专用模型）中新增的攻击面——指令后门攻击（instruction backdoor attacks）进行了研究。这类攻击中，攻击者将隐藏的恶意行为植入定制指令，而无需修改底
Private Direct Preference Optimization for LLM Alignment
直接偏好优化（DPO）已成为利用人类偏好数据对齐大语言模型（LLM）的标准方法。每个 DPO 样本包含一个提示（prompt）和一对候选模型回复，而回复之间的相对偏好往往反映标注者或终端用户的主观判断，可能揭示其敏感属性。现成的隐私保护方法
LoginTrap: Uncovering Task-Agnostic Phishing-Style Indirect Prompt Injection Attacks against LLM-based Web Agents
该论文研究基于大语言模型（LLM）的Web智能体所面临的一种新型安全威胁，即登录诱导型间接提示注入攻击。Web智能体通过观察网页内容并模拟用户执行浏览器操作来自动完成任务，而登录过程涉及凭证和敏感信息，是重要的身份验证边界。现有研究已表明恶