A robust autonomous state machine that ensures AI agents maintain continuous productivity and balanced attention across multiple long-term goals.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install goal-heartbeat
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 goal-heartbeat 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 Goal Heartbeat Engine?
The Goal Heartbeat Engine is a sophisticated state machine designed for AI agents operating within the Openclaw Skills framework. It transforms passive agents into proactive workers by implementing a heartbeat mechanism where every cycle must result in a tangible output. By rotating focus across various long-term objectives, the engine prevents path dependency and ensures that no critical task is left starved of attention.
This skill is essential for developers looking to build highly autonomous AI systems that can self-correct, prioritize, and deliver consistent value without constant human intervention. It serves as the rhythmic pulse of an agent's workspace, governing how attention is distributed and how progress is documented across complex workflows.
Goal Heartbeat Engine Use Cases
- Managing multiple long-term research projects simultaneously without human supervision.
- Maintaining continuous background task execution for autonomous AI agents in high-availability environments.
- Ensuring balanced progress across diverse engineering, marketing, and documentation goals.
- Implementing fail-safe mechanisms for agents to always have a productive task to execute even when triggers are absent.
- Scaling Openclaw Skills across multi-agent systems where attention allocation is critical.
How Goal Heartbeat Engine Works
- State Initialization: The agent begins by reading PULSE.md and GOALS.md to determine the current status, previous actions, and task hunger levels.
- Mode Selection: A dice roll mechanism, utilizing system timestamps, selects one of five modes: Miracle Mode, Random Dive, Relay Race, Supplement Advance, or Surprise Mode.
- Action Determination: The agent selects a specific action based on the chosen mode and checks for urgency overrides, such as forced relays for long-waiting tasks.
- Deep Execution: The agent performs deep work for a minimum of 5 minutes, involving multi-lingual searches and intensive source analysis to produce high-quality files.
- Output Delivery: Every cycle concludes with the generation of at least one tangible asset, such as research notes, scripts, or prototype demos.
- State Update: The agent closes the loop by updating the pulse trackers, goal progress, and internal memory before performing a final self-check to ensure compliance with quality standards.
Goal Heartbeat Engine Setup
To integrate this engine into your Openclaw Skills workflow, follow these setup steps:
- Initialize the required directory structure in your workspace.
- Create the core state files in the memory directory.
- Configure your reference documents to guide the agent's autonomous behavior.
# Create directory structure
mkdir -p workspace/memory references
# Initialize pulse and goal trackers
touch workspace/memory/PULSE.md
touch workspace/memory/GOALS.md
# Add reference protocols
touch references/action-protocols.md
touch references/seed-keywords.md
Goal Heartbeat Engine Data Schema & Taxonomy
The Goal Heartbeat Engine organizes data through a structured taxonomy in Markdown to ensure human-readability and agent-parsing efficiency:
| File | Purpose | Key Metadata |
|---|---|---|
| PULSE.md | State Machine Log | Last service time, goal rotation counter, relay status, and anti-duplication rules. |
| GOALS.md | Project Kanban | Milestones (G1-G4), subtask progress, and completion checkboxes. |
| references/ | Behavioral Logic | Seed keywords for random dives and protocol definitions for high-impact actions. |
Health tracking follows a strict hierarchy: STARVED (critical priority), COLD (needs attention), OK (normal rotation), and HOT (saturated).
Goal Heartbeat Engine Advanced Features
- Anti-duplication logic: Implements hard rules to prevent the agent from repeating the same target goal or output type more than three times consecutively.
- Urgency Overrides: High-priority triggers that automatically bypass the dice roll for critical user messages or tasks that have been starved of attention for over 24 hours.
- Emergency Idle Protection: A fail-safe list of productive tasks (e.g., Hacker News analysis, AI repository trials) ensuring the agent never enters an idle state.
- Intelligent Notification System: A rule-based alert system that balances update frequency with quiet hours to keep the user informed without causing notification fatigue within the Openclaw Skills environment.
SKILL.md
Loading
Related Openclaw Skills
Feishu Suite Skill
A comprehensive integration for Feishu that automates documents, spreadsheets, instant messaging, and organizational workflows.
Daily Evolution
An automated self-iteration engine that enables AI agents to learn from daily interactions, optimize toolchains, and refine their own core persona files.
GitHub Copilot CLI Workflow
A high-performance workflow for senior engineers to orchestrate GitHub Copilot CLI as a specialized team of AI agents.
NESP Marketing & Airdrop Engine
An automated growth engine designed to drive $NESP adoption through viral marketing and M2M education.
OpenClaw Toolbox
A comprehensive management suite for environment initialization, system maintenance, and automated multi-mode backups for AI agent workspaces.
Seedream Image Generation Skill
A professional AI image generation skill leveraging the Seedream 4.5 API to create, edit, and fuse high-quality visual content through intelligent prompt optimization.
Featured*
