server-agent
Universal production server monitoring, healing, and reporting skill for cmdop
server-agent
CMDOP Skill — install and use via CMDOP agent:
Universal production server monitoring, self-healing, and reporting skill for cmdop.
Watches over Docker containers, detects anomalies, heals what is safe to heal, scans for vulnerabilities, and writes structured Markdown reports — all driven by the cmdop agent.
Quick Start
Commands
| Command | Description |
|---|---|
check | Full health check: containers, disk, Redis. Auto-heals safe issues. |
scan | Security scan: trivy CVEs, open ports, auth logs. |
heal | Targeted remediation for a specific container. |
report | Write structured MD reports (daily, security). |
status | Quick server snapshot — no files written. |
All commands support --json for structured output and --dry-run to preview actions without executing them.
Environment Variables
| Variable | Required | Description |
|---|---|---|
REPORTS_DIR | yes | Absolute path to reports directory |
TELEGRAM_BOT_TOKEN | no | Bot token for Telegram notifications |
TELEGRAM_CHAT_ID | no | Target chat/channel ID |
Autonomy Model
- SAFE (executes immediately):
docker restart,docker system prune - ESCALATE (sends Telegram alert): image rebuild, config changes, containers with >2 restarts, OOM
- NEVER: database commands,
rm -rfoutside/tmp, firewall changes
Protected containers (never auto-restarted): shared-db-redis, traefik
Setup on Production Server
Two Modes of Operation
Scheduled monitoring (Python CLI)
Triggers call server-agent binary directly. No LLM in the check loop — fast, deterministic, cheap.
Interactive investigation (cmdop skill)
Use when you need the agent to reason about an incident:
The agent reads skill/readme.md as its system prompt and uses server-agent CLI as a tool to gather data and take action.
Reports Structure
$REPORTS_DIR/ index.md baseline/ports.json 2026-03-14/ 10-23-incident-reforms-django.md 14-05-security-scan.md daily-summary.md
Reviews
No reviews yet.