llm-censorship-steering

LLM 审查引导 本仓库包含论文 引导审查：揭示 LLM “思维”控制的表征向量（作者：Hannah Cyberey 和 David Evans）的代码实现。 我们提出了一种从 LLM 内部找到“引导向量”的方法，用于检测和控制模型输出中的审查程度。欢迎查看这篇 博客文章 以快速了解我们的工作。 试试我们的演示： 🐳 使用 DeepSeek-R1-Distill-Qwen-7B 引导 思想抑制 🦙 使用 Llama-3.1-8B-Instruct 引导 拒绝—顺从 注意： 两个演示都需要 Huggingface 账户。该演示托管在 Huggingface 的 ZeroGPU...

Securing Computer-Use Agents against Branch Steering Attacks

Modern Computer Use Agents CUAs directly interact with graphical user interfaces and execute third-party web tools, exposing them to indirect prompt injection across every rendered page and tool response. While the Dual-LLM pattern is the primary system-level architecture offering formal security...

CounterSteer: Suppressing Indirect Prompt Injection with Activation Steering

Indirect prompt injection makes an LLM agent treat untrusted retrieved text as instructions. We present CounterSteer, an inference-time defense that suppresses this behavior inside the model. Per model, a five-step recipe fits a residual-stream direction from paired episodes differing only in...

The like Trap: Multi-Stage Poisoning against Agents in Similarity-Based Recommendation Systems

With recent advancements in large language models LLMs and LLM-based agents, these agents are becoming increasingly autonomous and gaining broader access to act on users' behalf on the internet. However, the vulnerability of automated agents deployed on social media platforms e.g., for managing a...

Astra Linux – Vulnerability found in Linux 5.10, Linux 5.15

In the Linux kernel, the following vulnerabilities have been resolved: net/mlx5: Fixed the cleanup of steering rules The vport’s unicast, multicast, and multicast rules are not deleted during the teardown process when EEH occurs. Since the vport’s promiscuous settings unicast, unicast, and...

TrojanWorld: Backdooring World-Model Agents Via Imagination Steering

World models increasingly serve as the predictive core of model-based reinforcement learning agents, enabling them to simulate future dynamics and reason over imagined trajectories before acting. Their substantial training demands make pretrained world models attractive for distribution and reuse...

ECLIPSE: Self-Evolving Stealthy Prompt Injection Attack against Long-Horizon Agentic Systems

Recently, large language model LLM agents, such as Codex, Claude Code, and OpenClaw, have become capable of planning and executing long-horizon tasks through repeated tool calls. This capability also creates new opportunities for prompt injection. Existing attacks either place the malicious...

Bendix EC80 Brake ECU

ADVISORY SUMMARY Successful exploitation of these vulnerabilities could allow an attacker to cause the loss of ABS functions, steering assist, speedometer, shifting capabilities, or disable automatic traction control. 2. RECOMMENDED PRACTICES CISA recommends users take defensive measures to...

CVE-2026-72032

A flaw was found in the Linux kernel's net/mlx5 HWS Hardware Steering component. This vulnerability occurs when the hwsbwcmatchermove function attempts to resize a matcher, and the mlx5hwsmatcherresizesettarget operation fails. In such a scenario, a replacement matcher is allocated but not proper...

UBUNTU-CVE-2026-72006

In the Linux kernel, the following vulnerability has been resolved: net/mlx5: free mlx5stidxdata on final dealloc Workloads that repeatedly allocate and release mkeys carrying TPH steering-tag hints e.g. churning RDMA MRs leak one struct mlx5stidxdata per cycle; kmemleak flags it as unreferenced...

CVE-2026-72006

The CVE-2026-72006 entry relates to the Linux kernel (net/mlx5) where the backing mlx5_st_idx_data was not freed on final deallocation, causing a memory leak in workloads that repeatedly allocate/release mkeys with TPH steering-tag hints. The root cause is that mlx5_st_dealloc_index() erases the ...

PT-2026-72195

Name of the Vulnerable Software and Affected Versions Linux kernel affected versions not specified Description A memory leak occurs in the net/mlx5 component when workloads repeatedly allocate and release mkeys that carry TPH steering-tag hints, such as churning RDMA MRs. The issue arises because...

UBUNTU-CVE-2026-18536

Data::Entropy versions before 0.010 for Perl read remote entropy sources over plain HTTP. The Data::Entropy::RawSource::RandomOrg and Data::Entropy::RawSource::RandomnumbersInfo remote sources are accessed over plain HTTP. The Data::Entropy::RawSource::RandomOrg integrity check trivially matches...

Astra Linux – Vulnerability in Linux 6.12

In the Linux kernel, the following vulnerabilities have been resolved: net/mlx5e: TC, delete flows only for existing peers When deleting TC steering flows, iterate only over actual devcom peers instead of assuming all possible ports exist. This avoids touching non-existent peers and ensures that...

Astra Linux – Vulnerability in Linux 5.10

In the Linux kernel, the following vulnerabilities have been resolved: RDMA/mana: Fixed an error in the unwind process of manaibcreateqprss. Sashiko noted that the manaibcfgvportsteering function has a flaw; the standard destruction path fixes this issue...

Astra Linux – Vulnerability in Linux 5.10

In the Linux kernel, the following vulnerability has been resolved: RDMA/manaib: The behavior of disabling RX steering when an RSS QP is destroyed has been corrected. When an RSS QP is destroyed e.g., due to DPDK exit, the manaibDestroyQPrss function destroys the RX WQ objects, but it does not...

kiro-cybersecurity-skills

CyberSecurity Skills A collection of 15 security workflows co...

SUSE CVE-2026-46144

In the Linux kernel, the following vulnerability has been resolved: RDMA/mana: Fix error unwind in manaibcreateqprss Sashiko points out that manaibcfgvportsteering is leaked, the normal destroy path cleans it up...

CVE-2026-46144

A flaw was found in the Linux kernel's Remote Direct Memory Access RDMA subsystem, specifically within the mana driver. During an error unwind in the manaibcreateqprss function, a resource leak occurs where manaibcfgvportsteering is not properly cleaned up. This vulnerability could lead to resour...

AZL-88926 CVE-2026-46144 affecting package kernel for versions less than 6.6.141.1-1

In the Linux kernel, the following vulnerability has been resolved: RDMA/mana: Fix error unwind in manaibcreateqprss Sashiko points out that manaibcfgvportsteering is leaked, the normal destroy path cleans it up...