Runtime observability and policy enforcement for AI coding agents (www.oculisecurity.com)

Oculi has announced a new security layer designed to enhance the safety of AI coding agents by intercepting and enforcing policies on every tool call made by these agents—be it shell commands, file edits, or MCP calls. This solution operates seamlessly by integrating with development environments like Claude Code, Cursor, and Windsurf, ensuring that security checks occur without hindering performance. Users can specify permissions through simple YAML-based rules, allowing them to monitor and control actions such as blocking potentially destructive commands (e.g., "rm -rf") and restricting access to sensitive files like .env. This innovation is significant for the AI/ML community as it addresses critical security issues around AI tool usage, particularly in environments where sensitive data and execution commands may pose risks. With features like real-time telemetry, full audit trails, and enterprise-grade policy management, Oculi empowers developers and organizations to maintain control over their AI workflows effectively. By providing a centralized gateway and integrations across various tools, Oculi supports comprehensive security strategies, enabling safer implementation of AI coding agents in professional settings.