Now in research preview: Codex Security Review It takes a deeper look at GitHub pull requests for security issues, using repository context to surface actionable findings directly in the PR. See how to enable automatic reviews:

Build a plugin once and use it across compatible agent clients. Introducing Agent Plugins, an open standard developed with @awsdevelopers, @cursor_ai, @github, @code, and @vercel that packages Agent Skills and supports MCP server configurations in a shared format.

On July 23, ZAN and @MetaEraCN co-hosted an AMA titled "The Dark Forest in the AI Agent Era: When Agents Begin to Manage Assets." ZAN Global Head of Partnerships Alexandra Wang shared key insights on the convergence of AI Agents and Web3. AI Agents are shifting from conversation to action. Human roles are no longer about executing every transaction step by step, but about setting goals and defining authorization boundaries. One person working with multiple Agents may be able to accomplish what once required an entire team. Beyond trading, Agents are also becoming autonomous participants in on-chain economic activity, while DAO governance and data value return are accelerating toward real-world adoption. Security is a foundational infrastructure issue. Once Agents gain access to wallets and payment capabilities, the risks become multi-layered: execution failures, permission abuse, and accountability gaps. The industry should build fine-grained permission controls, Clear Signing mechanisms, and a zero-trust security framework. Infrastructure needs to be redesigned. Identity, payments, security, and compute trust all need to be restructured for the Agent economy. ZAN's goal is to lay a solid foundation for this new world, enabling Agents to have trusted identities, executable payments, and secure collaboration capabilities. When AI begins to hold assets and make transactions, infrastructure becomes the trust foundation of the new world. #AI# #AIAgents# #Web3# #Blockchain# #DigitalAssets# #Security#