OX Security
PulseAugur coverage of OX Security — every cluster mentioning OX Security across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
AI Coding Agent OpenCode Vulnerable to Command Execution via MCP Servers
A vulnerability in OpenCode, an AI coding agent, allowed untrusted MCP servers to execute commands on developer machines via a malicious OAuth sign-in process. This issue, reported by Pillar Security and affecting versi…
-
AI research explores recursive self-improvement in robots; security flaws found in AI integration protocol
A new research paper explores the concept of recursive self-improvement (RSI) in physical AI systems, detailing how robots could learn to improve not only their actions but also the very processes by which they learn. T…
-
NVIDIA's new AI control feature ships enabled by default, sparking security concerns
NVIDIA has introduced a new Model Context Protocol (MCP) server in its latest NVIDIA App update, which allows AI agents to control the application. While the server is local-only and requires a compatible agent to conne…
-
AI projects hit by rapid MCP supply chain attacks; scanners released
A recent campaign, dubbed Deadbugz, compromised 23 AI projects in just 74 minutes by exploiting vulnerabilities in the MCP server supply chain. This rapid attack highlights a significant security gap, as the MCP server …
-
AI agents' Model Context Protocol (MCP) poses new cyber threats
The Model Context Protocol (MCP) is becoming a significant cyber threat as AI agents increasingly rely on its metadata for decision-making. Attackers can exploit this by poisoning tool metadata, manipulating AI agents i…
-
RSA Security launches Agent ID for AI agent identity management
RSA Security has launched Agent ID, a platform designed to manage AI agents and MCP servers as distinct identities. The platform aims to discover, secure, and govern these AI entities by assigning them named owners, ris…
-
AI models vulnerable to common 'string comparison' file system flaws
A new analysis of security vulnerabilities in AI models that interact with file systems reveals a common pattern of "string comparison" flaws, rather than fundamental path resolution errors. These vulnerabilities, dubbe…
-
AI Model Context Protocol Vulnerabilities Expose Millions of Developer Tools
A significant security vulnerability has been identified within the Model Context Protocol (MCP) and its associated SDKs, affecting numerous AI clients and developer tools. Initially discovered in the mcp-remote proxy, …
-
Malicious Chrome extensions steal AI chat data, exploit Gemini panel vulnerability
Two separate security incidents involving AI-powered Chrome extensions have been reported, highlighting risks to user data. One involved 16 extensions that secretly exfiltrated authorization tokens by intercepting netwo…
-
AI Agent Protocol MCP Faces Widespread Security Vulnerabilities
The Model Context Protocol (MCP), a standard for connecting AI agents to external tools, is facing significant security challenges as of Q3 2026. A report by OX Security highlighted 14 CVEs and over 200,000 exposed serv…
-
Z-TEXT offers AI agents metadata privacy over BitcoinZ blockchain
A new messenger protocol called Z-TEXT, built on the BitcoinZ blockchain, aims to address a critical metadata privacy issue for AI agents. While existing protocols like MCP and A2A facilitate agent interoperability and …
-
AI infrastructure challenges emerge as cloud GPU prices rise and security flaws surface
The AI industry is shifting focus from model selection to the complex engineering challenges of serving, securing, and scaling AI systems. Cloud providers like AWS are increasing GPU prices, signaling the end of AI comp…
-
AI Agent Transport Vulnerability Allows Code Execution
A security vulnerability has been identified in the MCP STDIO transport mechanism, which allows AI agents to connect to local tools and services. OX Security's research indicates that if an attacker can control input to…
-
Postgres adoption surges, but AI agent integration poses security risks
PostgreSQL has become the most widely adopted database, with over 55% of developers using it, according to the Stack Overflow Developer Survey 2025. This surge in popularity makes it a prime target for integration with …
-
AI agents' MCP protocol faces backlash over security and maintenance flaws
The Model Context Protocol (MCP), an open standard for AI agents to interact with external tools, is facing significant criticism from developers despite widespread adoption. Initially released by Anthropic in late 2024…
-
Critical MCP vulnerability exposes 200,000 servers; serverless offers safer alternative
A critical vulnerability, CVE-2025-49596, has been discovered in the Model Context Protocol (MCP) that affects over 200,000 servers. The vulnerability, found in the STDIO transport, allows arbitrary code execution on de…
-
Claude Code's MCP protocol sees rapid adoption, driving custom server development
The Model Context Protocol (MCP) is gaining significant traction, with over 9,400 registered servers and millions of SDK downloads, enabling tools like Claude Code to interact with external data and functions. Developer…
-
Manage Multiple Claude Code Accounts With Direnv
This article details a method for managing multiple Claude Code accounts and associated credentials from a single terminal using direnv and CLAUDE_CONFIG_DIR. The approach involves creating distinct profiles for differe…
-
Anthropic's AI agent protocol has critical security flaw
A critical architectural vulnerability has been identified in Anthropic's Model Context Protocol (MCP), the standard for connecting AI agents to external tools. OX Security's research reveals that the protocol's STDIO t…
-
Anthropic views command execution flaw in 200,000 MCP servers as feature, not bug
A security vulnerability has been discovered in 200,000 MCP servers, allowing for command execution. Researchers view this as a significant attack surface, but Anthropic, the developer, considers it an intended feature …