AgentWatcher: A Rule-Based Prompt Injection Monitor

Large language models LLMs and their applications, such as agents, are highly vulnerable to prompt injection attacks. State-of-the-art prompt injection detection methods have the following limitations: 1 their effectiveness degrades significantly as context length increases, and 2 they lack...

[SECURITY] Fedora 44 Update: rust-resctl-bench-2.2.5-12.fc44

resctl-bench is a collection of whole-system benchmarks to evaluate resource control and hardware behaviors using realistic simulated workloads. Comprehensive resource control involves the whole system. Furthermore, testing resource control end-to-end requires scenarios involving realistic...

Towards Secure Retrieval-Augmented Generation: A Comprehensive Review of Threats, Defenses and Benchmarks

Retrieval-Augmented Generation RAG significantly mitigates the hallucinations and domain knowledge deficiency in large language models by incorporating external knowledge bases. However, the multi-module architecture of RAG introduces complex system-level security vulnerabilities. Guided by the R...

Security-By-Design for LLM-Based Code Generation: Leveraging Internal Representations for Concept-Driven Steering Mechanisms

Large Language Models LLMs show remarkable capabilities in understanding natural language and generating complex code. However, as practitioners adopt CodeLLMs for increasingly critical development tasks, research reveals that these models frequently generate functionally correct yet insecure cod...

CVE-2026-29075

Mesa is an open-source Python library for agent-based modeling, simulating complex systems and exploring emergent behaviors. In version 3.5.0 and prior, checking out of untrusted code in benchmarks.yml workflow may lead to code execution in privileged runner. This issue has been patched via commi...

CVE-2026-29075

Mesa is an open-source Python library for agent-based modeling, simulating complex systems and exploring emergent behaviors. In version 3.5.0 and prior, checking out of untrusted code in benchmarks.yml workflow may lead to code execution in privileged runner. This issue has been patched via commi...

EUVD-2026-10045

Mesa is an open-source Python library for agent-based modeling, simulating complex systems and exploring emergent behaviors. In version 3.5.0 and prior, checking out of untrusted code in benchmarks.yml workflow may lead to code execution in privileged runner. This issue has been patched via commi...

CVE-2026-29075

Mesa is an open-source Python library for agent-based modeling, simulating complex systems and exploring emergent behaviors. In version 3.5.0 and prior, checking out of untrusted code in benchmarks.yml workflow may lead to code execution in privileged runner. This issue has been patched via commi...

CVE-2026-29075 Mesa: Checking out of untrusted code in `benchmarks.yml` workflow may lead to code execution in privileged runner

Mesa is an open-source Python library for agent-based modeling, simulating complex systems and exploring emergent behaviors. In version 3.5.0 and prior, checking out of untrusted code in benchmarks.yml workflow may lead to code execution in privileged runner. This issue has been patched via commi...

CVE-2026-29075

Mesa CVE-2026-29075 affects the Mesa library (versions ≤ 3.5.0). The vulnerability arises when untrusted code is checked out in the benchmarks.yml workflow, potentially allowing code execution in a privileged runner. The issue has been patched via commit c35b8cd. Public-facing exploitation detail...

CVE-2026-29075 Mesa: Checking out of untrusted code in `benchmarks.yml` workflow may lead to code execution in privileged runner

Mesa is an open-source Python library for agent-based modeling, simulating complex systems and exploring emergent behaviors. In version 3.5.0 and prior, checking out of untrusted code in benchmarks.yml workflow may lead to code execution in privileged runner. This issue has been patched via commi...

The Agile FedRAMP Playbook, Part 4: Reactive Risk Management through Enriched Incident Response

In the final part of our series, we explore Reactive Risk Management. Discover how Wiz for U.S. Government transforms cloud detection and response to help satisfy FedRAMP Rev 5 IR controls and FedRAMP 20x detection benchmarks...

VMWare Aria Operations 8.x < 8.18.6 Multiple Vulnerabilities (VMSA-2026-0001)

According to its self-reported version, the instance of VMWare Aria Operations formerly vRealize Operations running on the remote web server is 8.x 8.18.6. It is, therefore, affected by the following: - VMware Aria Operations contains a command injection vulnerability. A malicious unauthenticated...

Jailbreak Foundry: From Papers to Runnable Attacks for Reproducible Benchmarking

Jailbreak techniques for large language models LLMs evolve faster than benchmarks, making robustness estimates stale and difficult to compare across papers due to drift in datasets, harnesses, and judging protocols. We introduce JAILBREAK FOUNDRY JBF, a system that addresses this gap via a...

CVE-2026-22720

VMware Aria Operations contains a stored cross-site scripting vulnerability. A malicious actor with privileges to create custom benchmarks may be able to inject script to perform administrative actions in VMware Aria Operations. To remediate CVE-2026-22720, apply the patches listed in the 'Fixed...

EUVD-2026-8709

VMware Aria Operations contains a stored cross-site scripting vulnerability. A malicious actor with privileges to create custom benchmarks may be able to inject script to perform administrative actions in VMware Aria Operations. To remediate CVE-2026-22720, apply the patches listed in the 'Fixed...

CVE-2026-23229 crypto: virtio - Add spinlock protection with virtqueue notification

In the Linux kernel, the following vulnerability has been resolved: crypto: virtio - Add spinlock protection with virtqueue notification When VM boots with one virtio-crypto PCI device and builtin backend, run openssl benchmark command with multiple processes, such as openssl speed -evp aes-128-c...

[SECURITY] Fedora 42 Update: rust-resctl-bench-2.2.5-10.fc42

resctl-bench is a collection of whole-system benchmarks to evaluate resource control and hardware behaviors using realistic simulated workloads. Comprehensive resource control involves the whole system. Furthermore, testing resource control end-to-end requires scenarios involving realistic...

Optimizing Agent Planning for Security and Autonomy

Indirect prompt injection attacks threaten AI agents that execute consequential actions, motivating deterministic system-level defenses. Such defenses can provably block unsafe actions by enforcing confidentiality and integrity policies, but currently appear costly: they reduce task completion...

[SECURITY] Fedora 43 Update: rust-resctl-bench-2.2.5-10.fc43

resctl-bench is a collection of whole-system benchmarks to evaluate resource control and hardware behaviors using realistic simulated workloads. Comprehensive resource control involves the whole system. Furthermore, testing resource control end-to-end requires scenarios involving realistic...