SLOPerating System

SLOP Lacks Optimal Performance(erating system) is a completely vibecoded operating system prototype that boots into a shell and is being built toward a POSIX-compliant microkernel architecture.

The goal of this project is simple: to see how far vibecoding has come, and if it can really make something on par with major operating systems.

Current Goal

• Keep a stable 32-bit kernel baseline that boots reliably on 486-class constraints.
• Move commands from shell builtins into real user binaries.
• Expand syscall/process foundations so userland programs become the default path.
• Preserve simple terminal UX now with GUI groundwork later.

Current Features

• GRUB Multiboot ISO boot path with a modular freestanding C kernel.
• Basic shell with some commands built in (help, clear, cd, heap, alloc, echo, panic, reboot, halt)
• COM1 serial logging/serial shell context
• Optional framebuffer console path in QEMU with VGA fallback
• Heap allocator
• Kernel panics
• Basic syscall ABI
• Very basic ELF32 support (with cat, touch, ls, pwd, mkdir, ps, and systeminfo as real ELF files)
• Basic in-memory VFS
• Block device registry + IDE/ATAPI CD + read-only ISO9660 file loading (/mount/cdrom)

Minimum Hardware Profile

• CPU: Intel 80486 (baseline target).
• RAM: 4 MB target baseline.
• Video: text-mode console baseline.

Notes:

• You cannot boot from a floppy (yet), so most real 486 systems probably can't load this
• Tested on v86 and QEMU so far (nothing else)

Build and Run

Requirements:

• nasm
• make (or mingw32-make on native Windows)
• i686-elf-gcc, i686-elf-ld
• grub-mkrescue (+ backend tooling such as xorriso)
• qemu-system-i386

Recommended on Windows: build from WSL Ubuntu.

make
make run

Useful targets:

• make run-lowmem (lower-RAM boot test path)
• make iso-fb / make run-fb (framebuffer QEMU path)
• make run QEMU_SERIAL_OPTS="-serial stdio" (serial shell on host terminal)
• make floppy / make run-floppy (legacy floppy boot path)

FAQ

Is this production-ready?

Yeah man!

What do you use?

ChatGPT Codex, Gemini, Claude, and other AI tools.

Why?

To see how far vibe coding has come.

Is this really a microkernel?

Directionally yes, currently no. The architecture is moving toward microkernel service boundaries, but most functionality is still in-kernel today.

Do you have any knowledge of coding, operating systems, etc?

Some, which helps keep experiments constrained and debuggable.