Qubes
Enables autonomous AI agents to manage Qubes-isolated environments with tag-scoped Admin API access and dom0-mediated trust boundaries.
This tool provides a highly secure framework for integrating autonomous AI agents into Qubes OS. It allows AI to interact with and manage designated Qubes VMs, granting real capabilities like provisioning qubes, building templates, and moving files, all while rigorously enforcing a dom0-mediated trust boundary. The operator's core system remains invisible and protected, with kernel-level isolation ensuring that AI agents operate within strictly defined, tag-scoped sandboxes, preventing unauthorized access or visibility into untagged resources.
主要功能
01Dom0-mediated wrappers (`qmcp.*`) for enforced trust boundaries
020 GitHub stars
03Tag-scoped Qubes Admin API access for AI agents
04Single-egress network sandbox with operator-chosen upstream
05Wrapped read operations to prevent existence leaks to AI agents
06Secure inter-qube file transfer between AI-managed qubes
使用案例
01Automating AI workflows within highly isolated Qubes OS environments
02Enabling AI agents to autonomously manage and provision Qubes VMs
03Facilitating AI-assisted security tasks like penetration testing in a controlled sandbox