agent-defibrillator

# Agent Defibrillator

Watchdog service that monitors your AI agent gateway and restarts it when it crashes.

## What It Does

- Checks gateway health every 10 minutes
- Detects crashes and stale processes
- Auto-restarts with cooldown protection
- Optional Discord notifications on restart
- Detects version mismatches after updates

## Install

**Recommended (review code first):**
```bash
git clone https://github.com/hazy2go/agent-defibrillator.git
cd agent-defibrillator
./install.sh
```

## Verify

```bash
launchctl list | grep defib
```

## Configure

Edit `~/.openclaw/scripts/defibrillator.sh`:

| Variable | Default | Description |
|----------|---------|-------------|
| `DEFIB_GATEWAY_LABEL` | `ai.openclaw.gateway` | launchd service label |
| `DEFIB_RETRY_DELAY` | `10` | Seconds between retries |
| `DEFIB_MAX_RETRIES` | `3` | Retries before restart |
| `DEFIB_COOLDOWN` | `300` | Seconds between restarts |
| `DISCORD_CHANNEL` | (empty) | Your channel ID for notifications |

## Commands

```bash
# View logs
tail -f ~/.openclaw/logs/defibrillator.log

# Manual check
~/.openclaw/scripts/defibrillator.sh

# Stop watchdog
launchctl bootout gui/$(id -u)/com.openclaw.defibrillator

# Restart watchdog
launchctl kickstart -k gui/$(id -u)/com.openclaw.defibrillator
```

## Uninstall

```bash
launchctl bootout gui/$(id -u)/com.openclaw.defibrillator
rm ~/Library/LaunchAgents/com.openclaw.defibrillator.plist
rm ~/.openclaw/scripts/defibrillator.sh
```

## Requirements

- macOS (uses launchd)
- AI agent running via launchd (OpenClaw, etc.)

# 🫀⚡ Agent Defibrillator

> *"CLEAR!"* — Every sysadmin, eventually

Your AI agent is amazing. It writes code, sends messages, automates your life. But sometimes... it just *dies*. Gateway stuck. Process frozen. Flatline.

And you're not home. Or you're asleep. Or you're touching grass (good for you).

**Agent Defibrillator** is a watchdog that monitors your AI agent gateway and automatically shocks it back to life when it crashes. No more coming home to a dead agent. No more "why didn't you reply for 6 hours?"

## How It Works

```
Every 10 minutes:
    │
    ├─ Agent alive?
    │       │
    │       └─ Version match? → 😴 (sleep)
    │               │
    │               └─ Stale version? → 🔄 Restart + 📢 Notify
    │
    └─ Agent dead?
           │
           ├─ Wait & retry (maybe it's just slow)
           │
           └─ Still dead after 3 checks?
                  │
                  └─ 🫀⚡ "CLEAR!" → Full restart + 📢 Notify
                         │
                         ├─ ✅ Pulse restored!
                         │
                         └─ 💀 Manual intervention needed
```

**Safety features:**
- 🔄 5-minute cooldown between shocks (no restart loops)
- 🔒 Lock file prevents concurrent runs
- 🧹 Kills orphaned processes before restart
- 📝 Only logs when something goes wrong (no spam)

**v2 features:**
- 🔍 **Version mismatch detection** — catches stale gateway processes running old code after npm updates
- 📢 **Discord notifications** — get pinged when your agent gets shocked back to life
- ⏰ 1-hour cooldown for version-based restarts (separate from crash cooldown)

## Quick Install

### Option 1: One-liner (recommended)
```bash
curl -fsSL https://raw.githubusercontent.com/hazy2go/agent-defibrillator/main/install.sh | bash
```

### Option 2: Clone & Install
```bash
git clone https://github.com/hazy2go/agent-defibrillator.git
cd agent-defibrillator
chmod +x install.sh
./install.sh
```

### Option 3: Ask Your Agent! 🤖

Just paste this prompt...