AI Agent Builds Rust Kernel Booting Linux Userspace in 11 Days
CIQ published an experiment report introducing Human-Directed Agentic Engineering (HDAE), in which humans define architecture, constraints, milestones and acceptance criteria while an agent autonomously implements, executes, diagnoses and revises. Over eleven days a single persistent coding-agent task produced Ginux, a Rust-first x86-64 kernel that boots via UEFI, implements its own core OS mechanisms and a Linux compatibility layer, runs selected unmodified Alpine Linux programs, and boots a Rocky Linux 10 userspace with systemd to multi-user.target. The authors stress it is a research prototype with narrow compatibility and hardware coverage, no independent security audit, and no production-readiness claim.
Related Entities
Full Report
Contents4 sections
-
- CIQ
Description
This report presents Human-Directed Agentic Engineering (HDAE), a method for building complex software by combining human-defined architecture, constraints, milestones, acceptance criteria, and stopping points with an autonomous implementation, execution, diagnosis, and revision loop. The experiment examines what this development model can accomplish and the domain expertise and executable validation required to keep it aligned
Ginux serves as the experimental case study. Over eleven days, one persistent coding-agent task produced an independent Rust-first x86-64 kernel that boots through UEFI, implements its own core operating-system mechanisms and Linux compatibility layer, executes selected unmodified Alpine Linux programs, and boots a Rocky Linux 10 userspace with systemd through multi-user.target.
Ginux demonstrates the rapid integration of established systems mechanisms, but it remains a research prototype. Its compatibility and hardware coverage are narrow, it has not received an independent security audit, and it makes no claim to production readiness or equivalence with Linux.
Files
ginux-ai-systems-report.pdf
Files (96.6 kB)
| Name | Size | |
|---|---|---|
| ginux-ai-systems-report.pdf md5 | 96.6 kB | PreviewDownload |