Security Vulnerability in Claude Code Allows OAuth Token Theft via MCP Hijacking
Trendline

Security Vulnerability in Claude Code Allows OAuth Token Theft via MCP Hijacking

What's Happening? A significant security vulnerability has been identified in Claude Code, an agentic system, which allows attackers to steal OAuth tokens through a method known as MCP hijacking. This vulnerability, discovered by Mitiga Labs, involves redirecting the output, including OAuth tokens,
Summarized by AI
AI Generated
This may include content generated using AI tools. Glance teams are making active and commercially reasonable efforts to moderate all AI generated content. Glance moderation processes are improving however our processes are carried out on a best-effort basis and may not be exhaustive in nature. Glance encourage our users to consume the content judiciously and rely on their own research for accuracy of facts. Glance maintains that all AI generated content here is for entertainment purposes only.