Semantic-Preserving Adversarial Attacks on LLMs: an Adaptive Greedy Binary Search Approach

Large Language Models LLMs increasingly rely on automatic prompt engineering in graphical user interfaces GUIs to refine user inputs and enhance response accuracy. However, the diversity of user requirements often leads to unintended misinterpretations, where automated optimizations distort...

Language of Network: a Generative Pre-Trained Model for Encrypted Traffic Comprehension

The increasing demand for privacy protection and security considerations leads to a significant rise in the proportion of encrypted network traffic. Since traffic content becomes unrecognizable after encryption, accurate analysis is challenging, making it difficult to classify applications and...

CVE-2025-5074

A vulnerability, which was classified as critical, was found in FreeFloat FTP Server 1.0. Affected is an unknown function of the component PROMPT Command Handler. The manipulation leads to buffer overflow. It is possible to launch the attack remotely. The exploit has been disclosed to the public...

CVE-2024-45989

Monica AI Assistant desktop application v2.3.0 is vulnerable to Exposure of Sensitive Information to an Unauthorized Actor. A prompt injection allows an attacker to modify chatbot answer with an unloaded image that exfiltrates the user's sensitive chat data of the current session to a malicious...

CVE-2024-7110

An issue was discovered in GitLab EE affecting all versions starting 17.0 to 17.1.6, 17.2 prior to 17.2.4, and 17.3 prior to 17.3.1 allows an attacker to execute arbitrary command in a victim's pipeline through prompt injection...

CVE-2024-7042

A vulnerability in the GraphCypherQAChain class of langchain-ai/langchainjs versions 0.2.5 and all versions with this class allows for prompt injection, leading to SQL injection. This vulnerability permits unauthorized data manipulation, data exfiltration, denial of service DoS by deleting all...

CVE-2024-25639

Khoj is an application that creates personal AI agents. The Khoj Obsidian, Desktop and Web clients inadequately sanitize the AI model's response and user inputs. This can trigger Cross Site Scripting XSS via Prompt Injection from untrusted documents either indexed by the user on Khoj or read by...

CVE-2024-25858

In Foxit PDF Reader before 2024.1 and PDF Editor before 2024.1, code execution via JavaScript could occur because of an unoptimized prompt message for users to review parameters of commands...

CVE-2024-23203

The issue was addressed with additional permissions checks. This issue is fixed in iOS 16.7.6 and iPadOS 16.7.6, iOS 17.3 and iPadOS 17.3, macOS Sonoma 14.3, macOS Ventura 13.6.5. A shortcut may be able to use sensitive data with certain actions without prompting the user...

CVE-2024-40836

A logic issue was addressed with improved checks. This issue is fixed in iOS 16.7.9 and iPadOS 16.7.9, iOS 17.6 and iPadOS 17.6, macOS Sonoma 14.6, watchOS 10.6. A shortcut may be able to use sensitive data with certain actions without prompting the user...

CVE-2024-40833

A logic issue was addressed with improved checks. This issue is fixed in iOS 16.7.9 and iPadOS 16.7.9, macOS Monterey 12.7.6, macOS Sonoma 14.6, macOS Ventura 13.6.8. A shortcut may be able to use sensitive data with certain actions without prompting the user...

CVE-2024-40835

A logic issue was addressed with improved checks. This issue is fixed in iOS 16.7.9 and iPadOS 16.7.9, iOS 17.6 and iPadOS 17.6, macOS Monterey 12.7.6, macOS Sonoma 14.6, macOS Ventura 13.6.8, watchOS 10.6. A shortcut may be able to use sensitive data with certain actions without prompting the us...

CVE-2024-5389

In lunary-ai/lunary version 1.2.13, an insufficient granularity of access control vulnerability allows users to create, update, get, and delete prompt variations for datasets not owned by their organization. This issue arises due to the application not properly validating the ownership of dataset...

CVE-2024-4099

An issue has been discovered in GitLab EE affecting all versions starting from 16.0 prior to 17.2.8, from 17.3 prior to 17.3.4, and from 17.4 prior to 17.4.1. An AI feature was found to read unsanitized content in a way that could have allowed an attacker to hide prompt injection...

CVE-2024-40787

This issue was addressed by adding an additional prompt for user consent. This issue is fixed in iOS 17.6 and iPadOS 17.6, macOS Monterey 12.7.6, macOS Sonoma 14.6, macOS Ventura 13.6.8, watchOS 10.6. A shortcut may be able to bypass Internet permission requirements...

CVE-2024-48144

A prompt injection vulnerability in the chatbox of Fusion Chat Chat AI Assistant Ask Me Anything v1.2.4.0 allows attackers to access and exfiltrate all previous and subsequent chat data between the user and the AI assistant via a crafted message...

CVE-2024-48139

A prompt injection vulnerability in the chatbox of Blackbox AI v1.3.95 allows attackers to access and exfiltrate all previous and subsequent chat data between the user and the AI assistant via a crafted message...

CVE-2024-48141

A prompt injection vulnerability in the chatbox of Zhipu AI CodeGeeX v2.17.0 allows attackers to access and exfiltrate all previous and subsequent chat data between the user and the AI assistant via a crafted message...

CVE-2024-5248

In lunary-ai/lunary version 1.2.5, an improper access control vulnerability exists due to a missing permission check in the GET /v1/users/me/org endpoint. The platform's role definitions restrict the Prompt Editor role to prompt management and project viewing/listing capabilities, explicitly...

CVE-2024-48140

A prompt injection vulnerability in the chatbox of Butterfly Effect Limited Monica Your AI Copilot powered by ChatGPT4 v6.3.0 allows attackers to access and exfiltrate all previous and subsequent chat data between the user and the AI assistant via a crafted message...