| Mobile Web

Search results for OAuth

AI & Enterprise

Unauthorised Claude token usage spreads after session hijacking, users say

Cases have emerged in which hackers hijacked login sessions for Anthropic’s Claude accounts and used Claude Code tokens without authorisation. Anthropic confirmed abnormal activity on some accounts, forced session terminations, invalidated server-side tokens and issued refunds to some users. A UK-based independent AI consultant reported unexplained usage increases on a paid plan and said detailed logs were unavailable. Similar complaints surfaced on Reddit and GitHub, including sudden usage spikes and billing changes.

AI & Enterprise

MCP protocol gets biggest upgrade since launch; what\'s changed

Open-source technology MCP, used to connect AI applications, has received its biggest update since launch. The update focuses on revamping how request metadata is handled, strengthening authentication security and adding an extension framework. The new version adopts a stateless core that embeds metadata directly in requests, making it easier to scale during traffic spikes and simplifying recovery after server outages. It also reduces risks such as OAuth mix-up attacks and includes built-in extension libraries.

AI & Enterprise

Okta, Google Cloud expand partnership to boost AI agents and browser security

Okta and Google Cloud are expanding their partnership with a focus on strengthening AI agent and web browser security, Techzine reported on June 16. The companies will protect AI agents through Auth0 for AI Agents and the Gemini Enterprise Agent Platform runtime, and secure web browsers with Chrome Enterprise. Okta said integrating with Google Gemini Enterprise brings enterprise-grade identity verification and access management for AI agents.