Heartbeat Checklist

What This Skill Does

The Heartbeat Checklist skill is an essential operational framework designed for OpenClaw agents to maintain long-term stability, performance, and context management. It functions as a structured routine, allowing the agent to perform self-audits 2-4 times daily. The skill covers a broad spectrum of responsibilities, ranging from technical system monitoring—such as checking resource-monitor timers and Hydra daemon status—to cognitive hygiene tasks like managing context window limits and distilling daily memory files. By implementing this protocol, users ensure that their OpenClaw agent remains responsive, secure, and proactively aligned with their goals, preventing the degradation commonly associated with long-running AI sessions.

Installation

To integrate the Heartbeat Checklist into your agent environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/felipetruman/heartbeat-checklist

Ensure that you have the necessary write permissions in your memory and system directories for the agent to execute cleanup and logging tasks effectively.

Use Cases

This skill is ideal for power users who keep their OpenClaw agent active for extended periods. It is highly recommended for developers managing multiple projects within a single context, users who require rigorous security posture monitoring, and those who want to offload the burden of routine system maintenance. It serves as a guardrail against memory bloat and ensures that critical project learnings are captured in MEMORY.md rather than being lost to session resets.

Example Prompts

1. "Run the heartbeat checklist now, specifically focusing on our current context percentage and memory cleanup."
2. "Perform the security scan and verify that the Hydra daemon is still active."
3. "Execute the full heartbeat protocol and update MEMORY.md with the key takeaways from today's work."

Tips & Limitations

Always heed the urgency warnings for context limits; hitting an 85% context threshold can lead to unpredictable behavior if not managed via a controlled /reset. While the skill automates much of the maintenance, human review of the distilled memory files is recommended once a week to ensure alignment. Note that system-level commands like systemctl require the agent to have appropriate local permissions to successfully re-enable background services. Finally, respect the strict rule regarding apis.md file integrity; do not authorize automated modifications unless strictly necessary for your integration workflow.