Decoding Latent Attack Surfaces in LLMs: Prompt Injection Via HTML in Web Summarization

Large Language Models LLMs are increasingly integrated into web-based systems for content summarization, yet their susceptibility to prompt injection attacks remains a pressing concern. In this study, we explore how non-visible HTML elements such as , aria-label, and alt attributes can be exploit...

Multimodal Prompt Injection Attacks: Risks and Defenses for Modern LLMs

Large Language Models LLMs have seen rapid adoption in recent years, with industries increasingly relying on them to maintain a competitive advantage. These models excel at interpreting user instructions and generating human-like responses, leading to their integration across diverse domains,...

CVE-2025-58372

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. Versions 3.25.23 and below contain a vulnerability where certain VS Code workspace configuration files .code-workspace are not protected in the same way as the .vscode folder. If the agent was configured to auto-appro...

CVE-2025-58372 Roo Code: Potential Remote Code Execution via .code-workspace

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. Versions 3.25.23 and below contain a vulnerability where certain VS Code workspace configuration files .code-workspace are not protected in the same way as the .vscode folder. If the agent was configured to auto-appro...

CVE-2025-58372 Roo Code: Potential Remote Code Execution via .code-workspace

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. Versions 3.25.23 and below contain a vulnerability where certain VS Code workspace configuration files .code-workspace are not protected in the same way as the .vscode folder. If the agent was configured to auto-appro...

CVE-2025-58372 Roo Code: Potential Remote Code Execution via .code-workspace

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. Versions 3.25.23 and below contain a vulnerability where certain VS Code workspace configuration files .code-workspace are not protected in the same way as the .vscode folder. If the agent was configured to auto-appro...

CVE-2025-58372

Roo Code CVE-2025-58372 affects versions ≤3.25.23 where certain VS Code workspace files (.code-workspace) aren’t protected like the .vscode folder. If auto-approve for file writes is enabled and prompts are manipulated (e.g., via prompt injection), an attacker could write malicious workspace sett...

CVE-2025-9959

Incomplete validation of dunder attributes allows an attacker to escape from the Local Python execution environment sandbox, enforced by smolagents. The attack requires a Prompt Injection in order to trick the agent to create malicious code...

PT-2025-36340

Name of the Vulnerable Software and Affected Versions: Roo Code versions 3.25.23 and below Description: Roo Code, an AI-powered autonomous coding agent, is susceptible to a flaw where VS Code workspace configuration files .code-workspace lack the same protection as files within the .vscode folder...

CVE-2025-58357 5ire Chat Message XSS Vulnerability Enables Remote Code Execution

5ire is a cross-platform desktop artificial intelligence assistant and model context protocol client. Version 0.13.2 contains a vulnerability in the chat page's script gadgets that enables content injection attacks through multiple vectors: malicious prompt injection pages, compromised MCP server...

CVE-2025-58357 5ire Chat Message XSS Vulnerability Enables Remote Code Execution

5ire is a cross-platform desktop artificial intelligence assistant and model context protocol client. Version 0.13.2 contains a vulnerability in the chat page's script gadgets that enables content injection attacks through multiple vectors: malicious prompt injection pages, compromised MCP server...

CVE-2025-58357

Summary: 5ire (desktop AI assistant) vulnerability in the chat page’s script gadgets leads to content injection via multiple vectors (malicious prompt injection pages, compromised MCP servers, exploited tool integrations). Affected version: 0.13.2. Impact: allows content injection; CVSS 3.1 base ...

CVE-2025-9959

Incomplete validation of dunder attributes allows an attacker to escape from the Local Python execution environment sandbox, enforced by smolagents. The attack requires a Prompt Injection in order to trick the agent to create malicious code...

CVE-2025-9959

CVE-2025-9959 concerns smolagents, a Python agents library. The root issue is incomplete validation of dunder attributes, enabling an attacker to escape the Local Python execution environment sandbox when a Prompt Injection is used to coax the agent into creating malicious code. Several connected...

CVE-2025-9959 Sandbox escape in smolagents Local Python execution environment via dunder attributes

Incomplete validation of dunder attributes allows an attacker to escape from the Local Python execution environment sandbox, enforced by smolagents. The attack requires a Prompt Injection in order to trick the agent to create malicious code...

CVE-2025-9959 Sandbox escape in smolagents Local Python execution environment via dunder attributes

Incomplete validation of dunder attributes allows an attacker to escape from the Local Python execution environment sandbox, enforced by smolagents. The attack requires a Prompt Injection in order to trick the agent to create malicious code...

CVE-2025-9959 Sandbox escape in smolagents Local Python execution environment via dunder attributes

Incomplete validation of dunder attributes allows an attacker to escape from the Local Python execution environment sandbox, enforced by smolagents. The attack requires a Prompt Injection in order to trick the agent to create malicious code...

Indirect Prompt Injection Attacks Against LLM Assistants

Really good research on practical attacks against LLM agents. "Invitation Is All You Need! Promptware Attacks Against LLM-Powered Assistants in Production Are Practical and Dangerous" Abstract: The growing integration of LLMs into applications has introduced new security risks, notably known as...

PT-2025-35805

Name of the Vulnerable Software and Affected Versions smolagents affected versions not specified Description Incomplete validation of dunder attributes allows an attacker to escape from the Local Python execution environment sandbox enforced by the software. The attack requires a Prompt Injection...

Cybersecurity AI: Hacking the AI Hackers Via Prompt Injection

We demonstrate how AI-powered cybersecurity tools can be turned against themselves through prompt injection attacks. Prompt injection is reminiscent of cross-site scripting XSS: malicious text is hidden within seemingly trusted content, and when the system processes it, that text is transformed...