Multi-Agent Task Automation with Role-Based .md Configurations & Automated Security Testing for SaaS

Feature Requests
1

I’d like to propose two powerful additions to Cursor AI:

  1. Role-Based Agents via .md Files – Allow users to create specialized agents for different tasks, where each agent’s role, scope, and behavior are defined by a simple Markdown (.md) configuration file. This would be similar to how Claude Code handles multi-agent roles, enabling users to switch between or orchestrate agents tailored for coding, documentation, debugging, research, etc. directly from chat window, without relying on background agents.

  2. Automated Security Testing for SaaS – Before publishing or deploying an application, Cursor should provide a built-in automated security testing feature. This would run checks for common vulnerabilities (e.g., OWASP Top 10, misconfigurations, dependency risks) and provide actionable insights, ensuring the product is secure before going live.

Together, these features would streamline multi-task workflows while making development safer and more production-ready.

1 Like