solomonneas/solo-mise

Solo-mise is a Python package that sets up a ready-to-use workspace for people running AI coding agents, tools like Claude Code, Codex, or OpenClaw, on real projects. The name comes from the kitchen term mise en place, which means having everything in its place before you start. Here, it means laying out all the rules, memory files, safety guards, and handoff mechanisms an agent needs before a session begins, so nothing important is scrambled or lost mid-run. The main problem it solves is memory and privacy. AI agents often forget context between sessions and can accidentally expose private details in public repositories. Solo-mise gives you a structured folder layout, memory cards, safety rules, identity files, and a shared handoff inbox, so agents can pass notes to each other and pick up where they left off. A publish guard also runs before any code leaves your machine, scanning commits for private hostnames, account IDs, or personal details that should not become public. You install it with a single command and run solo-mise init to bootstrap your workspace. You pick two settings: how much baseline material to install, a lightweight repo setup or a fuller workspace setup with memory cards and identity files, and which agent tool you use it with, Claude Code, Codex, OpenClaw, or Hermes. A doctor command then checks that everything is correctly wired before you rely on it. The package makes no network calls on its own. All setup happens locally on your filesystem, using templates packaged with the install. The only thing that touches the network is the pre-push hook, which runs the content guard scanner against your own commits before they leave your machine.