CSOonline

Prompt hijacking puts MCP-based AI workflows at risk

Model context protocol (MCP) gives IT teams a standardized way to connect large language models (LLMs) to tools and data sources when developing AI-based workflows. But security researchers warn that ...
Infosecurity-magazine.com

Prompt Injection Bugs Found in Official Anthropic Git MCP Server

Three security vulnerabilities in the official Git server for Anthropic's Model Context Protocol (MCP), mcp-server-git, have been identified by cybersecurity researchers. The flaws can be exploited ...

Lofty Launches MCP Server to Give Brokerages an Open Foundation for Their AI Strategy

Lofty, an award-winning AI innovator for the real estate industry, today announced the release of Lofty MCP Server, a new ...

Only .mcp.json was left outside the lock file — 91.8% of 110 MCP servers are unpinned, identified in 53 lines

However, the MCP servers defined in .mcp.json in the same repository sometimes ignore that lock file entirely.After ...
Cybernews

Thousands of LiteLLM instances exposed: 1 in 10 uses default password

Wiz, a Google-owned cybersecurity firm, scanned the internet and found thousands of exposed LiteLLM servers. Then it checked 3,074 of them and found 294 instances – roughly 1 in 10 – that require no ...

Eventtia Launches Native MCP Server to Help Event Teams Manage Events With AI

Eventtia today announced the launch of its native Model Context Protocol (MCP) server. The server connects Eventtia's event management platform with AI assistants, helping teams automate ...