Adversaries initiated a targeted reconnaissance campaign against vulnerable PraisonAI nodes less than four hours following the public disclosure of a critical security defect. An automated scanning entity identifying as CVE-Detector/1.0 launched offensives…

The Agentic SOC Era: Architecting Next-Generation Incident Response with Open Context Standardization

A critical MCP Toolbox vulnerability exposes enterprise databases. Learn how this security flaw allows session hijacking and how to fix it.

New research explores AI generated code vulnerabilities within MCP server repositories. Learn why expert domain verification remains critical.

Discover the proposed OpenAI token price reduction. Learn how the enterprise plans to cut costs to compete with Anthropic ahead of its upcoming IPO.

Claude Fable 5 is back online. Developers using Claude Code can access the model through their existing subscription until July 7. However, there is a significant catch. Usage is capped at 50 perc…

TL;DR A researcher found six SGLang vulnerabilities, and the worst allow unauthenticated remote code execution. SGLang serves large language models such as DeepSeek and Qwen. CERT/CC published the…

You’d never run an npm package you found in a random gist. Half the industry is doing exactly that with model weights and calling it…

I Audited 300 MCP-Related npm Packages. My Scanner Couldn’t Inspect Runtime Code in 217. So I created Driftward This is not a malware count. It is a warning about security reports that say “zero …

Recently, numerous Codex users encountered frequent model capacity errors during daily use. However, the selected model has not actually reached its true capacity limit. Instead, this error simply…

Introduction: The “One Token at a Time” Bottleneck

While the global technology sector passionately debates whether humanity should decelerate the development velocity of artificial intelligence, OpenAI, the vanguard of generative AI, has already c…

On September 11, 2026, a GitHub-reviewed advisory was published for the PyPI package mysql-mcp-server: its SSE transport validated neither…

Introduction

Researchers have uncovered a vast network of AI gateways that conceals the real users behind requests sent to Claude, OpenAI, Gemini, xAI, and other models. Team Cymru first counted 10,867 such se…

To combat unauthorized model distillation, Anthropic recently fortified its protective mechanisms. These defenses secure the reasoning outputs of Claude AI. Internal estimates show these novel saf…

Computer science researchers recently demonstrated a novel technique to conceal dormant functions within artificial neural networks. Even auditors with unrestricted access to model weights cannot …

An MCP server is effectively an API that agents call for people, and that gives it an unusual set of risks. The caller is a model that text…

Most teams are wiring Claude to their systems with duct tape, or not at all. Here is how to build an MCP connector that survives…

Update ZAP add-ons through an LLM client, compare three local PHP targets, and discover why a missing alert doesn’t mean you’re safe.