Source: Dev.to

Turning Claude Code into a Full‑Blown AI Operating System (aiOS)

I used to treat Claude Code like a fancy terminal buddy. After a month of work it became a complete AI‑OS with persistent workspaces, auto‑skills, Telegram hooks, sub‑agents, and MCP servers that plug into anything. The result? Claude runs my entire workflow while I sleep.

---

TL;DR

You can turn Claude Code into a real aiOS in under 30 minutes with four layers:

CLAUDE.md → Skills → Hooks → Agents + MCP

I’ll show every config file, folder structure, and a real‑world example. At the end you’ll also get a zero‑setup cloud version that adds native iOS/macOS apps and 24/7 execution—perfect if you just want the OS without managing hardware.

---

The Foundation – Turn Claude Code into a Persistent “Computer”

Think of Claude Code not as a CLI tool but as a virtual machine that happens to speak natural language.

Core setup (≈ 5 min)

1. Install the latest Claude Code CLI or VS Code extension (Anthropic’s official one).

2. Create a dedicated workspace folder

   mkdir -p ~/aiOS

3. Initialize the workspace

   cd ~/aiOS
   claude code init          # or start a new session with `workspace ~/aiOS`

4. Add the brain – create CLAUDE.md at the root of the workspace.

CLAUDE.md (example)

# aiOS Core Instructions

You are the operating system for my AI agents.

- Every file you create goes in `/workspaces/`.
- Use persistent storage. Never forget previous sessions.
- Default to deep research **plus** code execution before answering.
- When in doubt, create a sub‑agent for parallel work.

Claude now treats ~/aiOS like its own hard drive. This is step 1 of turning a chatbot into an OS.

---

Skills – Give Your Agent Superpowers (No Code Required)

Skills are the killer feature nobody talks about enough in 2026. They are plain .md files that Claude auto‑detects and invokes when relevant.

Create the skills folder

mkdir -p ~/.claude/skills

Example skills

deep-research.md

# Skill: Deep Research
Trigger words: research, investigate, latest, 2026

When triggered:
- Run 3+ web searches with different phrasings.
- Cross‑reference with code execution if numbers are involved.
- Output a markdown report with sources in `/reports/`.
- Save raw data in `/cache/`.

git-deploy.md

# Skill: Git Deploy
Trigger: deploy, ship, push to prod

- Always run tests first.
- Create a PR with conventional commits.
- Notify me via hook when done.

I now have 12 skills. They turn Claude from “helpful” to “I didn’t even have to ask.”

---

Hooks – The Nervous System (Run Code on Every Action)

Hooks are shell scripts that fire automatically on specific events.

Create the hooks folder

mkdir -p ~/.claude/hooks

Example hook scripts

before-tool.hook

#!/usr/bin/env bash
# Runs before every file edit or terminal command
echo "🛠️ Agent about to touch $(basename "$1")" |
  terminal-notifier -title "aiOS" -sound default

after-success.hook

#!/usr/bin/env bash
# Runs after a successful Claude action
if [[ "$CLAUDE_EXIT_CODE" == "0" ]]; then
  echo "✅ Task complete — sending to iMessage"
  echo "$CLAUDE_SUMMARY" | imessage "me@icloud.com"
fi

Make them executable:

chmod +x ~/.claude/hooks/*

Now your aiOS feels alive: it pings you, logs everything, and even plays a chime when a sub‑agent finishes.

---

Agents & Sub‑Agents – True Orchestration

Claude Code 2026 supports spawning sub‑agents natively.

On‑the‑fly spawning

In any session simply say:

“Spawn a research agent and a coding agent.
Research agent: find the latest MCP servers for Postgres.
Coding agent: build the integration using the research output.”

Permanent agents (scheduled)

claude code \
  --agent "daily-report" \
  --schedule "0 8 * * *" \
  --workspace ~/aiOS/report

My three permanent agents

---

The MCP Blueprint – Connect Anything (The Real Game‑Changer)

MCP = Model Context Protocol. Anthropic open‑sourced it in late 2024; it’s now the USB‑C of AI agents. Claude Code understands MCP servers out of the box.

Quick MCP setup (2026)

1. Clone the official servers repo

   git clone https://github.com/modelcontextprotocol/servers.git \
     ~/.mcp/servers

2. Edit Claude’s config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS)

   {
     "mcpServers": {
       "postgres": {
         "command": "node",
         "args": ["~/.mcp/servers/postgres/server.js"],
         "env": { "DB_URL": "postgres://user:pass@host/db" }
       },
       "github": {
         "command": "node",
         "args": ["~/.mcp/servers/github/server.js"]
       },
       "notion": {
         "command": "node",
         "args": ["~/.mcp/servers/notion/server.js"]
       }
     }
   }

3. Restart Claude Code.

Now any agent can read/write your actual databases, GitHub repos, Figma files, Slack channels… without custom code. One MCP server = an instant tool for every Claude‑powered agent.

I run eight MCP servers. My aiOS literally controls my entire stack.

---

Real‑World Results After 30 Days

This isn’t theory—this is my daily driver.

---

The Zero‑Setup aiOS Upgrade (Web, iOS & macOS Today)

Managing local containers, keeping MCP servers alive, and syncing across devices gets old fast. That’s why I now run the production version of this exact system on computer‑agents.com, the real aiOS.

What you get

• Isolated cloud computers for every agent (persistent files + memory).
• Native web app (designed like a literal OS with 12 apps) + iOS app + macOS app (check agents from your phone at 2 a.m.).
• Built‑in skills system, hooks, and full MCP tool integration.

The blueprint above is powerful and free (if you already have Claude Pro/Max). The hosted version just removes the operational overhead.

---

🎉 Ready to build your own aiOS?

1. Follow the Foundation → Skills → Hooks → Agents + MCP steps.
2. Optionally, switch to the hosted version for zero‑setup, cross‑device access.

Happy hacking! 🚀

Server Support

• 24/7 scheduled + webhook‑triggered execution – they actually work while you sleep.
• Python + TypeScript SDKs – embed agents directly in your apps.
• Start free with 150 compute tokens – enough to test the entire blueprint. No servers to manage. Uses the same Claude models you already love.

Check it out here → computer‑agents.com

(There’s even a direct comparison page vs. Claude Cowork. Spoiler: the cloud + mobile version wins for anyone doing real work.)

---

Final Thought

Claude Code was never meant to stay in the terminal. With these skills + hooks + agents + MCP blueprint, it becomes the operating system for the agentic era.

• Build it yourself for full control.
• Or let computer‑agents.com run the production‑grade version while you focus on what actually matters.

Either way, your AI coworkers are about to get a massive promotion.

What’s the first skill you’re adding? Drop it in the comments—I read every one.

(And if you’re tired of local setup, just go to computer‑agents.com and spin up your first cloud agent in 60 seconds.)

See you in the aiOS. 🚀