home / skills / openclaw / skills / tommy-monitoring-dashboard
This skill provides real-time zero-token Discord monitoring with auto-updating dashboards for OpenClaw, tracking activity, health, and performance.
npx playbooks add skill openclaw/skills --skill tommy-monitoring-dashboardReview the files below or copy the command above to add this skill to your agents.
---
name: live-monitoring-dashboard
description: "Zero-token real-time Discord monitoring dashboard for OpenClaw. Displays system health, cron jobs, sessions, and performance analytics via persistent Discord messages updated every 2 minutes with no LLM token cost."
metadata: { "openclaw": { "emoji": "π", "requires": { "bins": ["curl", "jq", "top", "df"] } } }
---
# Live Monitoring Dashboard
**Zero-Token Real-time Discord monitoring for your OpenClaw ecosystem**
## What it does
Creates a dedicated Discord channel with live-updating messages showing:
**π€ Activity Monitoring (Slice 1 & 2)**
- Active subagents and their current tasks
- Running and recent cron jobs with status
- OpenClaw activity and system processes
- Session count and activity overview
**π₯οΈ System Health (Slice 3)**
- Real-time CPU, memory, and disk usage
- System uptime and process monitoring
- Status indicators with color-coded alerts
**π Performance Analytics (Slice 4) β¨**
- Historical performance trend analysis
- Smart threshold-based alerting system
- Peak usage tracking and pattern detection
- Configurable alert thresholds
**Zero Token Cost:** Uses shell scripts + Discord CLI commands instead of expensive LLM calls for 100% cost efficiency.
Updates every 60 seconds with zero message spam - only edits existing messages.
## Installation
```bash
npm install
./install.sh
```
## Configuration
The skill will:
1. Create a dedicated monitoring channel (`#tommy-monitoring` by default)
2. Post initial dual dashboard messages (Activity + System Health)
3. Begin live zero-token updates automatically
4. Initialize performance analytics and alert thresholds
## Usage
Once installed, the monitoring runs automatically with these commands:
**Manual Performance Tracking:**
```bash
./scripts/performance-tracker.sh track # Log current metrics and check alerts
./scripts/performance-tracker.sh trends # Show performance trends
./scripts/performance-tracker.sh alerts # View recent alerts
./scripts/performance-tracker.sh cleanup # Clean old performance data
```
**Configuration:**
Performance thresholds can be customized in `config/performance-config.json`:
```json
{
"cpu_warning": 70,
"cpu_critical": 85,
"mem_warning": 75,
"mem_critical": 90,
"disk_warning": 80,
"disk_critical": 95,
"retention_days": 7
}
```
## Requirements
- Discord channel configured in OpenClaw
- `message` tool available
- Basic system utilities (`ps`, `top`, `uptime`, `df`)
- `jq` for JSON configuration parsing
- `bc` for mathematical calculations
## Complete Feature Set (All 4 Slices)
**β
Slice 1 & 2: Discord Activity Monitoring**
- OpenClaw session tracking and activity overview
- Cron job status monitoring (21+ jobs supported)
- Subagent detection and lifecycle tracking
- Process monitoring with real-time updates
**β
Slice 3: System Health Monitoring**
- Real-time CPU/memory/disk usage with smart indicators
- System uptime tracking and stability monitoring
- Critical events logging with 10-event circular buffer
- Status indicators: π’ Normal / π‘ Medium / β οΈ High / π¨ Alert
**β
Slice 4: Performance Analytics & Smart Alerts** β
- **Historical trend analysis:** CPU/memory patterns over time
- **Smart threshold alerting:** Configurable warning/critical levels
- **Peak usage tracking:** Daily maximums and averages
- **Data retention:** Configurable history (default 7 days)
- **Alert logging:** Comprehensive alert history with timestamps
- **Trend detection:** Simple directional analysis (βοΈβοΈβ)
**β
Zero Token Architecture:**
- **Shell-based updates:** Direct `openclaw message edit` CLI usage
- **No LLM inference:** Pure data collection and formatting
- **100% cost efficiency:** Eliminated 2.8M daily token waste
- **Sustainable operation:** Real-time monitoring with zero ongoing costs
## Architecture
**Update Cycle (60-second intervals):**
1. **Data Collection:** System metrics via shell commands (`top`, `df`, `uptime`)
2. **Performance Tracking:** Historical logging and trend analysis
3. **Alert Processing:** Threshold checking and alert generation
4. **Discord Updates:** Direct message editing via OpenClaw CLI
5. **Cleanup:** Automatic old data removal based on retention policy
**File Structure:**
```
live-monitoring-dashboard/
βββ scripts/
β βββ zero-token-dashboard.sh # Main dashboard updater
β βββ performance-tracker.sh # Slice 4 analytics
βββ config/
β βββ live-state.json # Runtime state
β βββ performance-config.json # Alert thresholds
βββ data/
βββ performance/ # Daily metrics (CSV)
βββ alerts.log # Alert history
```
## Performance
- **CPU Impact:** Minimal (shell commands only)
- **Memory Usage:** <1MB for data storage
- **Disk Usage:** ~100KB/day for metrics (auto-cleanup enabled)
- **Network:** Only outbound Discord API calls
- **Token Cost:** **ZERO** (no LLM inference required)
---
**π Ready for Production | Built for OpenClaw Community**
*Developed during trust window sessions March 5, 2026 - demonstrating zero-token operational excellence and partnership-level autonomous development.*This skill provides a zero-token real-time Discord monitoring dashboard for OpenClaw. It posts and persistently updates dashboard messages in a dedicated channel every 60β120 seconds to show activity, system health, and performance analytics. The design uses shell scripts and the OpenClaw CLI so there are no LLM token costs or runtime inference fees.
A background shell updater collects metrics (CPU, memory, disk, uptime, processes, cron jobs, sessions) and appends historical samples to lightweight CSV stores. An alert engine compares metrics to configurable thresholds and logs events. The updater edits persistent Discord messages via the OpenClaw message CLI so the channel shows live state without spamming new messages.
Does this use LLM tokens or external inference?
No. All collection, formatting, and Discord edits are shell-based and use the OpenClaw CLI, so there are no LLM token costs.
How often are updates posted?
By default messages are edited every 60 seconds; the interval can be adjusted in the updater script.
What happens to old performance data?
Data is retained based on retention_days in performance-config.json and old files are cleaned up automatically.