nateherkai/AIS-OS

Overview

nateherkai/AIS-OS is an AI Operating System starter kit for Claude Code — listed in the vault’s Resources page. It connects the LLM Wiki pattern to the broader “AI OS” framing that llm-as-os describes.

Key Facts

• Purpose: AI Operating System starter kit for Claude Code
• Role: scaffolding for an agent-driven OS over your workspace
• License: MIT
• Framework: Three Ms (Mindset, Method, Machine) + Four Cs (Context, Connections, Capabilities, Cadence)
• Listed alongside the Anthropic ecosystem in Resources

Features

Core Architecture

• AI OS starter kit — scaffolding for an agent-driven operating system over your workspace
• Three Ms framework — Mindset, Method, Machine as organizing principles
• Four Cs loop — Context → Connections → Capabilities → Cadence
• Litmus test — AIOS should observe a real-world event and produce output faster than operator alone
• Context-switching reduction — design goal: AIOS as default surface for thought work

Ingest & Processing

• /onboard — 7-question setup wizard, generates Day-1 file set + fills CLAUDE.md
• /level-up — recurring Three Ms interview, one run = one shipped artifact
• Voice samples — captured during onboarding for humanizer

Quality & Governance

• /audit — recurring Four-Cs gap report (form check)
• Decisions log — append-only decisions/log.md
• Archives — old stuff moved, never deleted
• Weekly ritual — /audit + /level-up cycle

MCP / Tools

• Context layer — CLAUDE.md loaded every session with business knowledge
• Connections layer — connections.md registry of reachable systems
• EXPANSIONS.md — roadmap for what to add as you grow

Integration

• Claude Code native — designed for Claude Code skill system
• Wiki as memory — LLM Wiki pattern serves as the OS memory layer

Data & Storage

• Markdown files — standard wiki output
• Session-based — context established per conversation

Relationships

• “implemented-by” llm-as-os — concrete implementation of the AI-OS framing
• “depends-on” claude-code — the agent it scaffolds
• “uses” llm-wiki — the knowledge layer of the OS
• “compatible-with” claude-skills — skills are part of the OS toolkit

Frameworks and Operating Loop

The repository is an MIT-licensed Claude Code starter kit with three recurring skills: /onboard establishes the initial context, /audit checks the Four Cs, and /level-up identifies one leverage opportunity. Its Three Ms are Mindset, Method, and Machine; its Four Cs are Context, Connections, Capabilities, and Cadence. The stated litmus test is that the AIOS should observe a real-world event and produce an output faster and more accurately than the operator would alone.

Implications: AIS-OS makes the AI-OS framing operational by sequencing context before connections and capabilities, and cadence last so automation is not applied to unproven workflows.

Implications

AIS-OS ties the wiki to the “OS de IA” vision the user has been exploring — it is the starter kit for running an agent OS over the workspace, with the wiki as memory.

External Links

• Repository: https://github.com/nateherkai/AIS-OS

Sources

^[wiki/concepts/resources.md]