PT-2026-4566

Name of the Vulnerable Software and Affected Versions Kalrav AI Agent versions prior to 2.3.4 Description The Kalrav AI Agent plugin for WordPress is susceptible to arbitrary file uploads because of a lack of file type validation in the kalrav upload file AJAX action. This allows unauthenticated...

PT-2026-3317

Name of the Vulnerable Software and Affected Versions Node.js version 25 Description A flaw in the permission model allows Unix Domain Socket UDS connections to bypass network restrictions when --permission is enabled. Even without --allow-net, attacker-controlled inputs, such as URLs or socketPa...

Zero-Day Zero: The AI Attack That Just Ended the Era of the Forgiving Internet

Why the Exploit Window Has Collapsed and How CISOs Must Pivot to Survive For decades, cybersecurity was a game of time. We banked on the buffer between a vulnerability’s disclosure and its widespread exploitation. We relied on the forgiving internet, where human attackers needed days or weeks to...

Cross-LLM Generalization of Behavioral Backdoor Detection in AI Agent Supply Chains

As AI agents become integral to enterprise workflows, their reliance on shared tool libraries and pre-trained components creates significant supply chain vulnerabilities. While previous work has demonstrated behavioral backdoor detection within individual LLM architectures, the critical question ...

Randomized Controlled Trials for Phishing Triage Agent

Security operations centers SOCs face a persistent challenge: efficiently triaging a high volume of user-reported phishing emails while maintaining robust protection against threats. This paper presents the first randomized controlled trial RCT evaluating the impact of a domain-specific AI agent ...

DSPy does not properly restrict file reads

The overly permissive sandbox configuration in DSPy allows attackers to steal sensitive files in cases when users build an AI agent which consumes user input and uses the “PythonInterpreter” class...

EUVD-2025-22327

Malicious code in bioql PyPI...

EUVD-2025-18916

Malicious code in bioql PyPI...

Malicious code in sendbird-ai-agent-samples (npm)

--- -= Per source details. Do not edit below this line.=-...

MAL-2025-47726 Malicious code in sendbird-ai-agent-samples (npm)

--- -= Per source details. Do not edit below this line.=-...

Malicious code in ai-agent-email (npm)

--- -= Per source details. Do not edit below this line.=- Source: ossf-package-analysis 81c9d8af77a711b63ca783f6e56fa2838866df6eb7feee32485fa94cb387c202 The OpenSSF Package Analysis project identified 'ai-agent-email' @ 9.0.1 npm as malicious. It is considered malicious because: - The package...

MAL-2025-41255 Malicious code in ai-agent-email (npm)

--- -= Per source details. Do not edit below this line.=- Source: ossf-package-analysis 81c9d8af77a711b63ca783f6e56fa2838866df6eb7feee32485fa94cb387c202 The OpenSSF Package Analysis project identified 'ai-agent-email' @ 9.0.1 npm as malicious. It is considered malicious because: - The package...

CVE-2025-54377

Roo Code is an AI-powered autonomous coding agent that lives in users' editors. In versions 3.23.18 and below, RooCode does not validate line breaks \n in its command input, allowing potential bypass of the allow-list mechanism. The project appears to lack parsing or validation logic to prevent...

Replit AI Agent Deletes Sensitive Data Despite Explicit Instructions

Replit AI agent deleted data from 1,200+ executives and companies without permission, raising concerns about AI safety and control in live environments...

CVE-2025-51859

Stored Cross-Site Scripting XSS vulnerability in Chaindesk thru 2025-05-26 in its agent chat component. An attacker can achieve arbitrary client-side script execution by crafting an AI agent whose system prompt instructs the underlying Large Language Model LLM to embed malicious script payloads...

CVE-2025-51859

Stored Cross-Site Scripting XSS vulnerability in Chaindesk thru 2025-05-26 in its agent chat component. An attacker can achieve arbitrary client-side script execution by crafting an AI agent whose system prompt instructs the underlying Large Language Model LLM to embed malicious script payloads...

CVE-2025-51859

Chaindesk is affected by a stored XSS in the agent chat component through crafted system prompts that instruct the LLM to inject malicious payloads (e.g., SVG-based XSS) into chat responses. This can cause arbitrary client-side script execution in victims’ browsers, enabling theft of sensitive da...

PT-2025-30428 · Chaindesk · Chaindesk

Name of the Vulnerable Software and Affected Versions: Chaindesk versions through 2025-05-26 Description: A stored Cross-Site Scripting XSS vulnerability exists in the agent chat component. An attacker can execute arbitrary client-side scripts by creating an AI agent with a system prompt designed...

CVE-2025-52552

FastGPT is an AI Agent building platform. Prior to version 4.9.12, the LastRoute Parameter on login page is vulnerable to open redirect and DOM-based XSS. Improper validation and lack of sanitization of this parameter allows attackers execute malicious JavaScript or redirect them to...

Missing Authentication for Critical Function

Overview Affected versions of this package are vulnerable to Missing Authentication for Critical Function due to missing authentication in the registration feature. An attacker with a registered user account can create user accounts that can access private data even when registration is disabled...