Discovering Universal Activation Directions for PII Leakage in Language Models

Modern language models exhibit rich internal structure, yet little is known about how privacy-sensitive behaviors, such as personally identifiable information PII leakage, are represented and modulated within their hidden states. We present UniLeak, a mechanistic-interpretability framework that...

PT-2026-23529

Name of the Vulnerable Software and Affected Versions OpenClaw versions prior to 2026.2.14 Description The Feishu extension in OpenClaw is susceptible to server-side request forgery SSRF. This allows attackers to retrieve content from attacker-controlled remote URLs without proper SSRF protection...

PT-2026-20963

Name of the Vulnerable Software and Affected Versions OpenClaw versions prior to 2026.2.15 Description OpenClaw embedded the current working directory workspace path into the agent system prompt without proper sanitization. An attacker could potentially exploit this by creating a directory with...

OpenClaw log poisoning (indirect prompt injection) via WebSocket headers

Summary In openclaw versions prior to 2026.2.13, OpenClaw logged certain WebSocket request headers including Origin and User-Agent without neutralization or length limits on the "closed before connect" path. If an unauthenticated client can reach the gateway and send crafted header values, those...

GHSA-G27F-9QJV-22PM OpenClaw log poisoning (indirect prompt injection) via WebSocket headers

Summary In openclaw versions prior to 2026.2.13, OpenClaw logged certain WebSocket request headers including Origin and User-Agent without neutralization or length limits on the "closed before connect" path. If an unauthenticated client can reach the gateway and send crafted header values, those...

GHSA-782P-5FR5-7FJ8 OpenClaw Affected by Remote Code Execution via System Prompt Injection in Slack Channel Descriptions

Summary When the Slack integration is enabled, Slack channel metadata topic/description could be incorporated into the model's system prompt. Impact Prompt injection is a documented risk for LLM-driven systems. This issue increased the injection surface by allowing untrusted Slack channel metadat...

OpenClaw Affected by Remote Code Execution via System Prompt Injection in Slack Channel Descriptions

Summary When the Slack integration is enabled, Slack channel metadata topic/description could be incorporated into the model's system prompt. Impact Prompt injection is a documented risk for LLM-driven systems. This issue increased the injection surface by allowing untrusted Slack channel metadat...

Microsoft Finds “Summarize with AI” Prompts Manipulating Chatbot Recommendations

New research from Microsoft has revealed that legitimate businesses are gaming artificial intelligence AI chatbots via the "Summarize with AI" button that's being increasingly placed on websites in ways that mirror classic search engine poisoning SEO. The new AI hijacking technique has been...

PT-2026-22438

Summary In openclaw versions prior to 2026.2.13, OpenClaw logged certain WebSocket request headers including Origin and User-Agent without neutralization or length limits on the "closed before connect" path. If an unauthenticated client can reach the gateway and send crafted header values, those...

PT-2026-20318

Name of the Vulnerable Software and Affected Versions OpenClaw versions prior to 2026.2.3 Description OpenClaw is a personal AI assistant. When the Slack integration is enabled, Slack channel metadata topic/description can be incorporated into the model's system prompt. This increases the injecti...

The Promptware Kill Chain

Attacks against modern generative artificial intelligence AI large language models LLMs pose a real threat. Yet discussions around these attacks and their potential defenses are dangerously myopic. The dominant narrative focuses on "prompt injection," a set of techniques to embed instructions int...

PT-2026-8013

Name of the Vulnerable Software and Affected Versions Cursor versions prior to 2.5 Description A sandbox escape allows for remote code execution RCE when the AI agent autonomously performs Git operations. A malicious actor can hide scripts within hidden Git hooks in nested bare repositories or us...

BIT-MOODLE-2025-67849 Moodle: moodle: cross-site scripting (xss) via improper sanitization of ai prompt responses

A flaw was found in Moodle. This cross-site scripting XSS vulnerability, caused by improper sanitization of AI prompt responses, allows attackers to inject malicious HTML or script into web pages. When other users view these compromised pages, their sessions could be stolen, or the user interface...

Prompt Injection Via Road Signs

Interesting research: "CHAI: Command Hijacking Against Embodied AI." Abstract: Embodied Artificial Intelligence AI promises to handle edge cases in robotic vehicle systems where data is scarce by using common-sense reasoning grounded in perception and action to generalize beyond training...

[SECURITY] Fedora 42 Update: rust-pretty-git-prompt-0.2.2-9.fc42

Your current git repository information inside a beautiful shell prompt...

Optimizing Agent Planning for Security and Autonomy

Indirect prompt injection attacks threaten AI agents that execute consequential actions, motivating deterministic system-level defenses. Such defenses can provably block unsafe actions by enforcing confidentiality and integrity policies, but currently appear costly: they reduce task completion...

Arbitrary File Read via Prompt Tag Source Validation Bypass in CreateModelVersion

The createmodelversion handler in mlflow/server/handlers.py uses a client-controlled tag to decide whether to skip source path validation. When a CreateModelVersion request includes the tag mlflow.prompt.isprompt, the helper ispromptrequest returns True, and the entire source validation block...

Manipulating AI memory for profit: The rise of AI Recommendation Poisoning

That helpful "Summarize with AI" button? It might be secretly manipulating what your AI recommends. Microsoft security researchers have discovered a growing trend of AI memory poisoning attacks used for promotional purposes, a technique we call AI Recommendation Poisoning. Companies are embedding...

Manipulating AI memory for profit: The rise of AI Recommendation Poisoning

That helpful "Summarize with AI" button? It might be secretly manipulating what your AI recommends. Microsoft security researchers have discovered a growing trend of AI memory poisoning attacks used for promotional purposes, a technique we call AI Recommendation Poisoning. Companies are embedding...

Improper Authentication

Overview @capgo/capacitor-native-biometric is a This plugin gives access to the native biometric apis for android and iOS Affected versions of this package are vulnerable to Improper Authentication via the onAuthenticationSucceeded function. An attacker can gain unauthorized access by hooking and...