# Claude Notify

> Configure desktop toast notifications for Claude Code hooks on Windows and macOS. Triggers when user asks to setup notifications, enable toast, configure alerts, or mentions "桌面通知", "通知配置", "toast notification setup".

- **Type:** Skill
- **Install:** `agentstack add skill-muxiaoming-my-claude-skills-claude-notify`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [muxiaoming](https://agentstack.voostack.com/s/muxiaoming)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [muxiaoming](https://github.com/muxiaoming)
- **Source:** https://github.com/muxiaoming/my-claude-skills/tree/master/skills/claude-notify

## Install

```sh
agentstack add skill-muxiaoming-my-claude-skills-claude-notify
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# Claude Notify

Configure desktop notifications for Claude Code events (Notification, PermissionDenied, Stop).

## Supported Platforms

| Platform | Method | Dependencies |
|----------|--------|-------------|
| Windows 10/11 | BurntToast PowerShell module | `Install-Module BurntToast -Scope CurrentUser` |
| macOS | osascript (built-in) | None |

## Setup Steps

### 1. Detect Platform

```bash
# Check OS
if [[ "$OSTYPE" == "msys" || "$OSTYPE" == "cygwin" || "$OSTYPE" == "win32" ]]; then
  PLATFORM="windows"
elif [[ "$(uname)" == "Darwin" ]]; then
  PLATFORM="macos"
fi
```

### 2. Install Dependencies (Windows only)

```powershell
Install-Module BurntToast -Scope CurrentUser -Force
```

### 3. Deploy Scripts

Copy the scripts from this skill's `scripts/` directory to `~/.claude/`:

- **Windows**: `scripts/toast.ps1` → `~/.claude/toast.ps1`
- **macOS**: `scripts/notify.sh` → `~/.claude/notify.sh` (chmod +x)

### 4. Configure Hooks

Update `~/.claude/settings.json` with the appropriate hooks for the detected platform.

**Windows hooks:**
```json
"hooks": {
  "Notification": [{"hooks": [{"command": "powershell -ExecutionPolicy Bypass -File 'C:\\Users\\\\.claude\\toast.ps1' -Type Notification", "type": "command"}]}],
  "PermissionDenied": [{"hooks": [{"command": "powershell -ExecutionPolicy Bypass -File 'C:\\Users\\\\.claude\\toast.ps1' -Type PermissionDenied", "type": "command"}]}],
  "Stop": [{"hooks": [{"command": "powershell -ExecutionPolicy Bypass -File 'C:\\Users\\\\.claude\\toast.ps1' -Type Stop", "type": "command"}]}]
}
```

**macOS hooks:**
```json
"hooks": {
  "Notification": [{"hooks": [{"command": "bash ~/.claude/notify.sh Notification", "type": "command"}]}],
  "PermissionDenied": [{"hooks": [{"command": "bash ~/.claude/notify.sh PermissionDenied", "type": "command"}]}],
  "Stop": [{"hooks": [{"command": "bash ~/.claude/notify.sh Stop", "type": "command"}]}]
}
```

### 5. Verify

Ask the user to restart Claude Code, then trigger a notification to confirm it works.

## Post-Setup

After successful configuration, inform the user:

> Notification setup complete!
>
> You can optionally delete this skill to save ~30 tokens of context. It will NOT affect your notifications.
> Delete this skill?

If user confirms deletion, remove the `~/.claude/skills/claude-notify/` directory.

## Notification Content

All platforms show the same info:
- **Title**: `[ProjectName] Claude Waiting / Permission Needed / Claude Done`
- **Body**: Dir, Branch, Model (3 lines)

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [muxiaoming](https://github.com/muxiaoming)
- **Source:** [muxiaoming/my-claude-skills](https://github.com/muxiaoming/my-claude-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-muxiaoming-my-claude-skills-claude-notify
- Seller: https://agentstack.voostack.com/s/muxiaoming
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.