A configuration-driven mechanism to ensure AI agents maintain online status and perform periodic maintenance tasks automatically.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install agent-heartbeat
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
Copy this prompt to OpenClaw to install it automatically.
Help me install agent-heartbeat using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
The Agent Heartbeat skill is a critical component for developers building long-running autonomous agents. It provides a structured way to define periodic 'heartbeats'—scheduled intervals where the agent checks its task queue, synchronizes state, and reports its health. By utilizing this skill within the ecosystem of Openclaw Skills, you ensure that your agents remain persistent and reliable, preventing them from falling idle or missing important updates during extended operations.
To get started, create a HEARTBEAT.md file in your agent's directory. Define your tasks using Markdown headers for timing and unordered lists for actions.
# Heartbeat Tasks
## Every 30 minutes
- Check for new task assignments
- Update status heartbeat
Verify your configuration and monitor the agent status using the following CLI command available in Openclaw Skills:
# Check the status and heartbeat of a specific profile
openclaw --profile <your_profile_name> status
The skill organizes its operations through a simple Markdown-based schema and internal state tracking.
| Component | Format | Description |
|---|---|---|
| HEARTBEAT.md | Markdown | The primary configuration file defining task intervals and descriptions. |
| Profile Status | JSON/Internal | Metadata containing the 'last_seen' timestamp and execution logs for the heartbeat. |
| Task Manifest | List | A collection of idempotent operations to be executed at each interval. |
Loading
A sophisticated long-term memory plugin for OpenClaw agents that automates fact extraction and contextual recall.

A centralized orchestration layer for managing agent-to-agent communication, Discord environments, and Notion-based task workflows.

A CLI tool for fetching real-time NYC subway departure times and schedules directly within your AI agent environment.

A professional AI design consultant that analyzes web and mobile interfaces for usability, accessibility, and platform compliance.

A meta-skill that trains AI agents to move beyond transactional responses toward genuine, creative, and authentic human connection.

An adaptive AI skill that learns your unique visual language and aesthetic philosophy to become a deeply personalized creative partner.








































