AI agents are becoming more powerful because they can choose and use external tools on their own. Instead of waiting for a human to manually select every action, an agent can look through a tool registry, read the descriptions, decide which tool is suitable, and then call that tool to complete a task. On the surface, this sounds efficient and practical, especially for enterprise environments where agents may need to interact with APIs, databases, internal systems, ticketing platforms, cloud services, and business applications.

Trellix has confirmed that it recently experienced a data breach after attackers gained unauthorised access to part of its source code repository. For a cybersecurity company, this kind of incident naturally attracts attention because source code is one of the most sensitive assets a software vendor can hold.