Cursor has rolled out several new features targeting development teams and AI workflow users. The "Hooks" feature, now in beta, allows users to script and customize the behavior of AI Agents during run time, offering granular control such as usage auditing, command blocking, and secret redaction. Early adopters have highlighted the flexibility this introduces for advanced automation and compliance needs.
"Team rules" now enable organizations to define global behavioral policies from a central dashboard, extending even to the Bugbot integration. This move addresses longstanding requests for uniformity across repositories in large teams.
Cursor 1.7 is now available!
— Cursor (@cursor_ai) September 30, 2025
As you type a prompt, suggestions now appear. Press Tab to accept.
Also new: custom hooks, deeplinks, team-wide rules, menubar support, and more. pic.twitter.com/mu7c9AP28R
The introduction of shareable deeplinks for prompts streamlines collaboration and onboarding, particularly for technical documentation and internal knowledge sharing. The system's "sandboxed terminals" now execute commands in isolated environments, limiting internet access and containing potential risks. This security measure is especially relevant for teams on allowlist mode, and developers retain fallback options if sandboxing interferes with operations.
Monitoring Agent status from the menubar and direct image file support for Agents further align the product with real-world team workflows, making it easier to track processes and utilize visual data.
Cursor, known for its developer-centric AI tools, continues to focus on secure collaboration and automation within coding environments. These updates reflect the company’s commitment to giving engineering teams more control, security, and customization. The features are currently available to users with varying access—some are in public beta, while others may be limited to teams or specific platform users.
