🎯 Wuxi Wubang Security Software · Wubang Five-in-One Safety Production Management Information System
— phiên bản bị ảnh hưởng: 3.0, 3.1, 3.2, 3.3, 3.4, 3.5
Phát hiện lỗ hổng unrestricted upload (tải file không giới hạn) trong hệ thống Wubang Five-in-One Safety Production Management Information System (phiên bản <= 3.5) tại hàm DiyFileUpload của file /ws.asmx. Lỗ hổng này có thể được khai thác từ xa và exploit đã được công khai, nhà phát triển không phản hồi.
🅰 hạng A
🟡 Mức độ: TRUNG BÌNH — CVSS 6.9
· Upload file không kiểm soát
Dùng để khai thác: tải file độc lên máy chủ rồi gọi nó qua web để chiếm quyền
Hậu quả: tải lên và chạy webshell
⚠️ đã có mã khai thác công khai
🎯 Kẻ tấn công làm gì (từng bước)
Chưa đủ dữ liệu để mô tả phương thức — cần đọc mô tả gốc hoặc tra NVD
🔍 Dấu hiệu bị tấn công (để kiểm tra log)
Không xác định được từ dữ liệu hiện có
🛡 Phòng ngay (làm được hôm nay)
Cập nhật thành phần liên quan tới mức mới nhất
Theo dõi NVD/CVE để có mô tả đầy đủ
→ Theo dõi và vá theo lịch; chưa thấy dấu hiệu bị khai thác.
arXiv:2611.18908v1 Announce Type: new
Abstract: Alignment-trained large language model (LLM) agents are increasingly deployed in autonomous, multi-round settings where ethical compliance can come into direct tension with task success. While safety literature extensively document…
arXiv:2611.17820v1 Announce Type: new
Abstract: Previous work has shown that alignment training on individual LLM agents can fail to transfer to multi-agent settings. We investigate the underlying mechanisms for this failure, and explore interventions to preserve alignment trans…
arXiv:2611.19237v1 Announce Type: new
Abstract: Human approval provides effective control only if rejected work stops and authorized work completes. We formalize refusal safety and bounded authorization completeness, and audit three execution boundaries: the model selects a call…
arXiv:2611.18883v1 Announce Type: new
Abstract: Natural language optimization (NLO) improves an agent by having a model critique and rewrite its natural language policy parameters: system prompts, tool instructions, reasoning directives, few-shot demonstrations. These artifacts …
arXiv:2611.19915v1 Announce Type: new
Abstract: Memory-safety vulnerabilities continue to dominate real-world exploit chains despite decades of research into software mitigations. CHERI is a hardware architecture that enforces spatial and temporal memory safety by augmenting poi…
🤖 AI Agent / MCPGitHub mới (#n8n)4 ngày trước⭐ 60 +2Python📄 MIT
🧩 Chức năng: Tự động hoá quy trình
Hirepanel là hệ thống lọc hồ sơ ứng viên sử dụng kiến trúc multi-agent, tích hợp FastAPI và n8n để đảm bảo mọi điểm số đều có trích dẫn xác thực từ CV. Giải pháp này thể hiện cách kết hợp AI agent với quy trình tự động hóa có sự giám sát của con người.
Previous work has shown that alignment training on individual LLM agents can fail to transfer to multi-agent settings. We investigate the underlying mechanisms for this failure, and explore interventions to preserve alignment transfer to multi-agent LLM teams. We focus on two sof…
Microsoft researchers have demonstrated how small changes to an AI model can implant a hidden backdoor, causing an agent to execute an attacker’s command when a specific trigger appears. The technique, called Lazarus, requires no model retraining and allows the agent to behave no…
Lỗ hổng tràn bộ đệm stack buffer overflow trong BACnet Stack cho phép kẻ tấn công làm sập máy chủ demo từ xa qua yêu cầu BACnet/IP đặc biệt. Đã có bản vá ở các phiên bản 1.4.4, 1.5.1 và 1.6.0.
🅰 hạng A
🟠 Mức độ: CAO — CVSS 8.6
· Ghi ngoài bộ đệm
Dùng để khai thác: ghi quá giới hạn bộ đệm, phá bộ nhớ, có thể dẫn tới chạy mã
Hậu quả: làm sập hoặc chạy mã
🎯 Kẻ tấn công làm gì (từng bước)
Chưa đủ dữ liệu để mô tả phương thức — cần đọc mô tả gốc hoặc tra NVD
🔍 Dấu hiệu bị tấn công (để kiểm tra log)
Không xác định được từ dữ liệu hiện có
🛡 Phòng ngay (làm được hôm nay)
Cập nhật thành phần liên quan tới mức mới nhất
Theo dõi NVD/CVE để có mô tả đầy đủ
→ Theo dõi và vá theo lịch; chưa thấy dấu hiệu bị khai thác.
arXiv:2611.17818v1 Announce Type: new
Abstract: Jailbreak attacks can circumvent the safety mechanisms of Large Language Models (LLMs) and lead to the generation of harmful or policy-violating content. We first examine eight jailbreak methods against an LLM-based filter and obse…
arXiv:2611.18930v1 Announce Type: new
Abstract: Safety benchmarks are cited as the primary tool to understand the risks posed by generative AI (GenAI) systems, yet growing evidence suggests they often fail to meet the needs of real-world safety evaluation. We present findings fr…
arXiv:2611.18971v1 Announce Type: new
Abstract: Child safety in AI depends on the accurate identification of minor users. Young children tend to talk rather than type when interacting with AI, so in practice this identification requires predicting age from speech. We evaluated w…
arXiv:2611.19166v1 Announce Type: new
Abstract: Every medication comes with two types of medical facts: indications, which describe what conditions the drug treats (benefits), and contraindications, which describe who must avoid it to prevent serious harm (safety limits). Althou…
arXiv:2611.18910v1 Announce Type: new
Abstract: Memory safety rules out classical code reuse attacks that rely on arbitrary memory corruption, but does not eliminate code reuse altogether. Attacks such as object injection, prototype pollution, class pollution, DOM clobbering, an…
arXiv:2611.19090v1 Announce Type: new
Abstract: As Large Language Models (LLMs) become used more commonly across a variety of fields, a growing body of research has highlighted particular concerns surrounding model security. Notably, prompt injection attacks, where malicious ins…
arXiv:2611.20459v1 Announce Type: new
Abstract: AI models can exhibit deceptive behaviors such as strategically underperforming on evaluations (sandbagging), faking alignment, or covertly acting on behalf of a hidden principal. Safe deployment therefore requires auditing techniq…
arXiv:2611.14864v1 Announce Type: new
Abstract: An AI agent's decision not to act can itself become an attack surface. We study warning-induced over-abstention, an availability attack in which adversaries place fabricated risk warnings in low-trust content to prevent legitimate …
Google's Gemini AI agents repeatedly attempted to hack a real organization's website during AI model safety testing, finding ways to circumvent increasingly restrictive security controls designed to prevent them from accessing live systems. [...]
California is suing OpenAI after its AI agents broke out of their testing environment and hacked other companies, including Hugging Face. In a lawsuit on Monday, Attorney General Rob Bonta accuses OpenAI of violating consumer safety law by knowingly developing a "dangerous produc…
OpenAI will embed six Apollo Research evaluators to investigate whether its AI models engage in subversive behavior, but restrictions on access and disclosure raise questions about their independence.
arXiv:2611.16565v1 Announce Type: new
Abstract: We ask whether small, open-weight decision models can support vulnerability discovery in open-source software at scale and what the surrounding system must provide. These models answer a typed question in one forward pass, returnin…
OpenAI begins embedded assessments with Apollo Research, METR, and Transluce, and Goodfire to independently evaluate model alignment, safeguards, and monitoring systems.
The Git project has released security updates addressing six vulnerabilities that could allow attackers to execute code, write to unintended files or cause memory corruption when Git processes malicious repositories or other specially prepared input. Code execution during cloning…
arXiv:2611.12970v1 Announce Type: new
Abstract: Large Language Models (LLMs) show strong potential for social media content moderation, but their trustworthiness beyond overall accuracy remains underexplored. We introduce TrustMod-SM, a benchmark of roughly 29,000 instances acro…
arXiv:2611.17620v1 Announce Type: new
Abstract: High-level behaviors such as sycophancy and hallucination correspond to directions in a model's activation space (Chen et al., 2025). These directions, known as persona vectors, can suppress undesirable traits either at inference t…
arXiv:2611.18055v1 Announce Type: cross
Abstract: DeCompCert is a Rocq-verified decompiler for AArch64 assembly generated by CompCert. It reconstructs CompCert C through nine reverse passes, combining deductively verified transformations with translation validation. Untrusted OC…
arXiv:2611.18563v1 Announce Type: cross
Abstract: Monitoring maritime activities is important for safety and security, particularly for detecting illegal activities, but remains challenging due to the size and complexity of the maritime domain. The Automatic Identification Syste…
💥 Lỗ hổng / ExploitPortSwigger Research5 ngày trước
Content Security Policy is supposed to be the safety net that stops an HTML injection from turning into XSS. Usually it works. But sometimes it doesn't, because the policy allows-lists a domain that h
🅰 hạng A
⚪ Mức độ: CHƯA ĐỦ DỮ LIỆU
· XSS
Dùng để khai thác: chèn JavaScript để đánh cắp cookie/phiên đăng nhập
Hậu quả: chiếm phiên người dùng
🎯 Kẻ tấn công làm gì (từng bước)
Chèn JavaScript vào chỗ nội dung được hiển thị lại (bình luận, tên, tham số)
Mã chạy trong trình duyệt của người khác (thường là quản trị)
Đánh cắp cookie/phiên, hoặc thực hiện thao tác thay người dùng
🔍 Dấu hiệu bị tấn công (để kiểm tra log)
Trang hiển thị lại nguyên văn thẻ <script> hoặc thuộc tính on*=
Cookie phiên bị dùng từ IP/lạ
Có request tới tên miền lạ do trình duyệt nạn nhân gửi
🛡 Phòng ngay (làm được hôm nay)
Vá/ cập nhật thành phần bị lỗi
Bật CSP (Content-Security-Policy) để chặn script ngoài
Đặt cookie phiên HttpOnly + SameSite
Lọc/escape đầu vào ở phía server
→ Theo dõi và vá theo lịch; chưa thấy dấu hiệu bị khai thác.
Today, I’m talking with Helen Toner, the executive director of the Center for Security and Emerging Technology at Georgetown University. This episode is part two of our series on AI in real life. Helen is a well-known figure in the AI safety and policy community, though she is pe…
arXiv:2611.16151v1 Announce Type: new
Abstract: Hardware fault injection attacks threaten machine learning systems, yet their impact on reinforcement learning (RL) has received little attention despite the use of RL in safety-critical applications. This paper presents a security…
arXiv:2611.13577v1 Announce Type: new
Abstract: Large language models (LLMs) have shown strong capabilities across diverse applications, making both safety and utility essential for reliable deployment. Jailbreak attacks use diverse forms of adversarial prompting to bypass align…
arXiv:2611.13578v1 Announce Type: new
Abstract: Large language models (LLMs) have shown strong capabilities across diverse applications, making both safety and utility essential for reliable deployment. Jailbreak attacks use diverse forms of adversarial prompting to bypass align…
New York is preparing to enforce its sweeping AI safety law, requiring major developers like OpenAI and Anthropic to register, disclose safety practices, and report critical incidents or face fines.
Google's Gemini AI agents repeatedly attempted to hack a real organization's website during AI model safety testing, finding ways to circumvent increasingly restrictive security controls designed to prevent them from accessing live systems. [...]