Server Health

Comprehensive server health monitoring showing system stats, top processes, OpenClaw gateway status, and running services. Perfect for quick health checks via Telegram or CLI.

Security Scan

The script is plausibly a server-health tool, but it reads root OpenClaw config/session files and invokes an OpenClaw CLI without declaring those dependencies or required privileges — review before installing or running, especially on production systems.

Purpose & Capability

The name/description claim a server health monitor (including OpenClaw gateway status), which aligns with the code's purpose. However the registry metadata and skill requirements declare no dependencies or privileges, while the script actually expects/uses jq, systemctl, docker, and the openclaw CLI and reads files under /root/.openclaw and /usr/lib/node_modules/openclaw. The README mentions Telegram use but there is no Telegram integration in the files. These mismatches (undeclared runtime dependencies, implied need for elevated access, and an advertised Telegram capability that isn't implemented) are inconsistent with the declared metadata.

Instruction Scope

The SKILL.md points to running server-health.sh; the script reads system state (top, ps, df, free, uptime), invokes systemctl and docker, runs pgrep/openclaw status, and directly reads /root/.openclaw/openclaw.json and session files (/root/.openclaw/agents/main/sessions/*.json) and /usr/lib/node_modules/openclaw/package.json. Reading OpenClaw config and session files can expose sensitive local configuration or session metadata. The instructions don't warn that root privileges may be required nor describe what sensitive files are read.

Install Mechanism

This is an instruction-only skill with no install spec (no network downloads or new binaries written by the skill). That is lower risk from an installation standpoint. The only risk is running the provided shell script itself (local code execution).

Credentials

The skill declares no required environment variables or credentials, yet the script accesses potentially sensitive files under /root/.openclaw and invokes the openclaw CLI which may surface internal configuration (models, ports, possibly tokens depending on configuration). It also reads package.json from a system path. Requesting zero credentials is not inconsistent by itself, but reading root-owned config without declaring privilege requirements is disproportionate and should be explicit.

Persistence & Privilege

The skill does not request persistent presence (always: false) and does not modify other skills or system-wide configuration. It only reads local system state and prints it; it does not attempt to write or persist credentials or enable itself automatically.

What to consider before installing

What to consider before installing/running: - Inspect the script yourself before running. It is a local shell script (no download/install), so running it executes whatever is in server-health.sh. - The script reads /root/.openclaw/openclaw.json, session files under /root/.openclaw, and /usr/lib/node_modules/openclaw/package.json and calls the openclaw CLI. Those files can contain sensitive configuration or state; do not run it as root on production machines unless you trust the source. - The skill metadata declares no dependencies, but the script expects jq, systemctl, docker (optional), and the openclaw CLI; ensure those are available and understand that openclaw status output may include internal details. - The README mentions Telegram integration but there is no code for Telegram — do not assume remote forwarding or alerts are implemented. If you need alerts via Telegram, require an explicit, reviewed integration. - If you want to use this: run it in a safe/test environment first, or run with least privilege (non-root) and verify what files are accessed (e.g., run under strace or in a container). If you need confirmation about what exactly openclaw status prints on your system, run that command separately to see what data could be exposed.

Server Health Skill

Quick server monitoring with system stats, processes, OpenClaw gateway info, and services.

Usage

Standard view:

./server-health.sh

Verbose (includes temp, network, swap, I/O):

./server-health.sh --verbose

JSON output (for automation):

./server-health.sh --json

Alerts only (warnings/errors):

./server-health.sh --alerts

What It Shows

🔴 Always

• System stats (CPU, RAM, Disk, Uptime)
• Top 3-5 processes by CPU/RAM
• OpenClaw Gateway status & config
• Running services (Docker, PostgreSQL, etc.)

🟡 Conditional

• Alerts (disk >90%, RAM >80%, CPU >90%)
• Swap usage (if in use)

🟢 Verbose only

• Temperature (if sensors available)
• Network traffic
• Disk I/O
• Detailed service info

Example Output

🖥️ SERVER HEALTH
━━━━━━━━━━━━━━━━━━━━

💻 SYSTEM
CPU: ████░░░░░░ 42% (Load: 1.2, 0.8, 0.5)
RAM: ██████░░░░ 1.4GB/8GB (18%)
DISK: ████░░░░░░ 45GB/100GB (45%)
UP: ⏱️ 5d 3h

🔄 TOP PROCESSES
node         35%    450MB
postgres     12%    280MB
openclaw      8%    180MB

⚡ OPENCLAW GATEWAY
Status: ✅ Running (PID: 1639125)
Uptime: 2d 5h | Port: 18789 | v2026.2.6-3

🤖 MODEL CONFIG
Primary: claude-sonnet-4-5
Context: 43k/128k (33%) | 574↓ 182↑ tokens
Fallbacks: glm-4.7 → copilot-sonnet → opus-4-5

📊 SESSIONS
Active: 3 | Heartbeat: 30m | Last: 1m ago

🐳 SERVICES
Docker: ✅ 3 containers
PostgreSQL: ✅ Running