Model Context Protocol
Coverage of Model Context Protocol in the Nexus archive.
- Snowflake buys Natoma to help freeze out rogue agents
Snowflake acquires Natoma, a startup developing a gateway for managing AI agent permissions across enterprise applications, to enhance its 'agentic control plane' strategy. The acquisition aims to enable secure AI agent actions within Snowflake's products while enforcing enterprise security policies.
- Millions of AI agents imperiled by critical vulnerability in open source package
A critical vulnerability in the Starlette open-source framework, used by millions of AI agents and tools, allows hackers to breach servers and steal sensitive data. Starlette, with 325 million weekly downloads, underpins frameworks like FastAPI and enables access to external systems via the Model Context Protocol (MCP), making stolen credentials highly valuable.
- Anthropic won't own MCP 'design flaw' putting 200K servers at risk, researcher says
A design flaw in Anthropic's Model Context Protocol (MCP) could expose up to 200,000 servers to potential takeover, according to security researchers. The flaw, described as either a bug or intentional but poor design, allows for unauthorized access or control.