Agentic-SOC-Simulation

🛡️ Agentic SOC Parallel Simulation ASPS: Autonomous Security Operations Parallel Simulation Center Redefining the Future of Security Operations: Full-Stack SOC Parallel Simulation, Multi-Agent Collaboration, and Data-Driven Autonomous Defense 🇨🇳 中文文档 Chinese Version This project is a cutting-edge...

CVE-2026-95183

An issue in the context-compaction feature of DeepSeek Harness dsh commit c291e796 allows attackers to execute a prompt injection attack via crafted external content...

CVE-2026-95182

An untrusted search path in DeepSeek Harness dsh commit c291e796 allows attackers to execute arbitrary code via supplying and opening crafted executable files in place of legitimate content...

CVE-2026-95183

An issue in the context-compaction feature of DeepSeek Harness dsh commit c291e796 allows attackers to execute a prompt injection attack via crafted external content...

EUVD-2026-102906

An issue in the context-compaction feature of DeepSeek Harness dsh commit c291e796 allows attackers to execute a prompt injection attack via crafted external content...

CVE-2026-95182

An untrusted search path in DeepSeek Harness dsh commit c291e796 allows attackers to execute arbitrary code via supplying and opening crafted executable files in place of legitimate content...

CVE-2026-95183

An issue in the context-compaction feature of DeepSeek Harness (dsh) , specifically within commit c291e796 , allows attackers to perform a prompt injection attack by providing crafted external content. The vulnerability is associated with CWE-94 (Improper Control of Generation of Code). No specif...

PT-2026-115624

Name of the Vulnerable Software and Affected Versions DeepSeek Harness dsh version c291e796 Description An untrusted search path allows attackers to execute arbitrary code by supplying and opening crafted executable files instead of legitimate content. Recommendations At the moment, there is no...

EUVD-2026-102903

An untrusted search path in DeepSeek Harness dsh commit c291e796 allows attackers to execute arbitrary code via supplying and opening crafted executable files in place of legitimate content...

CVE-2026-95182

A vulnerability in DeepSeek Harness (dsh) , specifically in commit c291e796 , involves an untrusted search path (CWE-426). This flaw allows an attacker to achieve arbitrary code execution by supplying and opening crafted executable files that replace legitimate content. The impact is rated as CRI...

CVE-2026-52394

DeepSeek V4 is vulnerable to Incorrect Access Control via a crafted input...

llm-censorship-steering

LLM Censorship Steering This repository contains code implementation for Steering the CensorShip: Uncovering Representation Vectors for LLM "Thought" Control by Hannah Cyberey and David Evans. We introduce a method that finds "steering vectors" from LLM internals for detecting and controlling the...

CVE-2026-52394

The DeepSeek V4 model is vulnerable to Incorrect Access Control when processing a crafted input. While specific internal component details or root cause files are not provided in the documents, the vulnerability allows for unauthorized access. Based on the SSVC metrics, the issue is considered au...

CVE-2026-108789

A vulnerability was identified in deepseek-ai deepseek-harness up to 0.1.1-rc.2. This affects the function agent of the component dsh-workflow-worker-thread. The manipulation leads to enforcement of behavioral workflow. The attack is possible to be carried out remotely. The exploit is publicly...

CVE-2026-108788

A vulnerability was determined in deepseek-ai deepseek-harness up to 0.1.1-rc.2. The impacted element is the function spawn of the component Sandbox. Executing a manipulation of the argument Profile can lead to sandbox issue. It is possible to launch the attack on the local host. The exploit has...

EUVD-2026-97158

A vulnerability was identified in deepseek-ai deepseek-harness up to 0.1.1-rc.2. This affects the function agent of the component dsh-workflow-worker-thread. The manipulation leads to enforcement of behavioral workflow. The attack is possible to be carried out remotely. The exploit is publicly...

CVE-2026-108789

A vulnerability exists in deepseek-ai deepseek-harness up to version 0.1.1-rc.2 . The flaw specifically affects the agent function within the dsh-workflow-worker-thread component. The root cause involves improper enforcement of behavioral workflow and business logic errors (CWE-841, CWE-840), whi...

CVE-2026-108789 deepseek-ai deepseek-harness dsh-workflow-worker-thread agent behavioral workflow

A vulnerability was identified in deepseek-ai deepseek-harness up to 0.1.1-rc.2. This affects the function agent of the component dsh-workflow-worker-thread. The manipulation leads to enforcement of behavioral workflow. The attack is possible to be carried out remotely. The exploit is publicly...

CVE-2026-108789 deepseek-ai deepseek-harness dsh-workflow-worker-thread agent behavioral workflow

A vulnerability was identified in deepseek-ai deepseek-harness up to 0.1.1-rc.2. This affects the function agent of the component dsh-workflow-worker-thread. The manipulation leads to enforcement of behavioral workflow. The attack is possible to be carried out remotely. The exploit is publicly...

CVE-2026-108788

A vulnerability exists in deepseek-ai deepseek-harness up to version 0.1.1-rc.2 . The issue is located within the spawn function of the Sandbox component. By manipulating the Profile argument, an attacker can cause a sandbox issue. This attack can be launched on the local host . While the exploit...