A production-grade operational toolkit for health monitoring, automated service recovery, and intelligent alerting for autonomous agent environments.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-ops-kit
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install agent-ops-kit using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Agent Ops Kit?
The Agent Ops Kit is a comprehensive operational suite designed to provide reliability and observability for autonomous agent systems. As an essential part of the Openclaw Skills ecosystem, it offers a robust framework for monitoring URLs, ports, and processes, ensuring that your agents remain operational 24/7. It moves beyond simple logging by providing an integrated system for health tracking and proactive maintenance.
This kit is specifically engineered for developers running Openclaw Skills in production who need enterprise-level features like atomic task management, daily uptime reporting, and automated self-healing. By utilizing a lightweight, file-based architecture with locking mechanisms, it provides a stable foundation for agent management without the overhead of heavy infrastructure.
Agent Ops Kit Use Cases
- Continuous health monitoring of web endpoints and API servers to ensure maximum availability.
- Automated recovery of crashed background workers or local server processes using configurable restart commands.
- Real-time alerting via Telegram with built-in rate limiting to prevent notification storms during outages.
- Long-term uptime tracking and daily performance digest generation for Openclaw Skills infrastructure.
- Lightweight task board management for tracking operational issues and investigation status through a CLI-friendly JSON format.
How Agent Ops Kit Works
- The orchestrator initializes the environment by reading a centralized services.json configuration file.
- Scheduled health checks are performed across defined URLs, TCP ports, and system processes.
- Upon detecting a failure, the recovery engine evaluates the restart logic and executes shell-based recovery commands within defined cooldown periods.
- Alerting modules verify the incident against the rate-limiting state before dispatching formatted Markdown notifications to a Telegram bot.
- Performance data is appended to daily metrics files for later analysis and uptime reporting.
- The task board system uses atomic writes and file locking to manage operational tasks without data corruption.
Agent Ops Kit Setup
Begin by creating the local directory structure for logs, metrics, and configuration:
mkdir -p ~/.agent-ops/{config,logs,metrics,scripts,state}
Next, define your monitored services in ~/.agent-ops/config/services.json. Ensure you include your Telegram bot token and chat ID for alerting. Once configured, deploy the health-check scripts and set up a periodic trigger using crontab or systemd to keep your Openclaw Skills monitored:
# Example crontab entry for 5-minute intervals
*/5 * * * * ~/.agent-ops/scripts/agent-ops-orchestrator.sh >> ~/.agent-ops/logs/cron.log 2>&1
Agent Ops Kit Data Schema & Taxonomy
The Agent Ops Kit organizes data using a clean, file-based taxonomy for portability:
| Data Type | Format | Location |
|---|---|---|
| Service Config | JSON | ~/.agent-ops/config/services.json |
| Performance Metrics | JSONL | ~/.agent-ops/metrics/YYYY-MM-DD.jsonl |
| Task Board | JSON | ~/.agent-ops/state/board.json |
| Alert State | JSON | ~/.agent-ops/state/alert-rate.json |
| Operational Logs | Plain Text | ~/.agent-ops/logs/ |
The metrics schema captures name, status (up/down), latency_ms, and an ISO-formatted timestamp for every check.
Agent Ops Kit Advanced Features
- Smart auto-recovery engine with configurable max-restart attempts and cooldown cycles.
- Atomic, lock-protected task board I/O to prevent race conditions during concurrent agent access.
- Intelligent Telegram alerting with topic-based deduplication and recovery notifications.
- Automated disk space monitoring and SSL certificate expiry tracking recipes.
- Maintenance utilities for log rotation and historical metrics cleanup to optimize storage for Openclaw Skills users.
SKILL.md
Loading
Related Openclaw Skills
Apple Notes Skill
A professional macOS automation skill for managing Apple Notes through search, creation, and structured content updates.
Video Editor Online Free
A cloud-powered AI video editing skill that enables users to trim, edit, and export 1080p MP4 files without local software installation.
JobHound
JobHound is a comprehensive AI-driven career companion designed to guide candidates through the entire recruitment lifecycle, from initial job discovery to final salary negotiation.
PV-52 Crisis Consultant
A high-pressure risk management AI agent designed to provide structural stability and strategic decision-making during organizational crises.
AI Hairstyle Changer OpenAPI Skill
A high-performance AI tool that modifies a person's hairstyle based on reference photos or text descriptions using the WeShop OpenAPI.
Adaptive Reasoning
An intelligent preprocessing skill that evaluates request complexity to dynamically adjust an agent's reasoning depth and token usage.