Anthropic Expands Claude Access for Vetted Cyber Teams as Glasswing Finds 129,000 Flaws

Anthropic on Tuesday said it's expanding a program that allows vetted cybersecurity professionals to test its advanced artificial intelligence AI models with reduced safeguards and blocking classifiers, as the company claimed its Project Glasswing initiative uncovered at least 129,000 verified...

Enhancing AI security through global AI red teaming

In this article 1. Building a global alliance 2. What the research focuses on 3. Why this matters Most AI safety testing still happens inside the walls of individual organizations. That has resulted in a fundamental disconnect: many of the highest-risk failure modes in modern AI systems require...

Microsoft Open-Sources RAMPART and Clarity to Secure AI Agents During Development

Microsoft has unveiled two new open-source tools called RAMPART and Clarity to assist developers in better testing the security of artificial intelligence AI agents. RAMPART, short for Risk Assessment and Measurement Platform for Agentic Red Teaming, functions as a Pytest-native safety and securi...

Introducing RAMPART and Clarity: Open source tools to bring safety into Agent development workflow

In this article 1. Why we are investing in this 2. RAMPART: Continuous safety testing for agentic AI 3. Clarity: Helping check software engineering assumptions 4. RAMPART and Clarity available now The AI systems shipping inside enterprises today are fundamentally different from the ones we were...

AEGIS: No Tool Call Left Unchecked -- a Pre-Execution Firewall and Audit Layer for AI Agents

AI agents increasingly act through external tools: they query databases, execute shell commands, read and write files, and send network requests. Yet in most current agent stacks, model-generated tool calls are handed to the execution layer with no framework-agnostic control point in between...

Grok apologizes for creating image of young girls in “sexualized attire”

Another AI system designed to be powerful and engaging ends up illustrating how guardrails routinely fail when development speed and feature races outrun safety controls. In a post on X, AI chatbot Grok confirmed that it generated an image of young girls in “sexualized attire.” The potential...

SQLi_AI_defence

SQLiAIdefence A small model ba...

A week in security (December 4 – December 10)

Last week on Malwarebytes Labs: Meta’s Purple Llama wants to test safety risks in AI models US government is snooping on people via phone push notifications, says senator Android phones can be taken over remotely – update when you can How IT teams can conduct a vulnerability assessment for...

Struts2 exploits tool Devmode version released with the source code-the vulnerabilities and early warning-the black bar safety net

! Disclaimer: This tool is for security testing purposes, the prohibition of the illegal use. Please pay attention and check the tool Safety. When Struts2 turn on devMode mode, will lead to a serious remote code execution vulnerability. If the WebService to start a permission is the highest...

Ajax hacking (Monyer)-vulnerability warning-the black bar safety net

Author: dream light In the tenth period published the Ajax hacking, the users give me feedback onXSSthe technical aspects of the problem mainly by the following several. For what in Ajax hacking, withXSS for? It with the traditionalXSSwhat's the difference? They each have what the pros and cons o...