A lightweight SQLite-backed coordination bus for sharing state, alerts, and tasks across multiple agent sessions.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install clankerhive
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 clankerhive 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 ClankerHive?
ClankerHive is a specialized shared context storage solution designed to facilitate seamless communication and coordination between multiple AI agent sessions. By utilizing a single SQLite database in WAL mode, it provides a robust and reliable alternative to ad-hoc JSON state files, ensuring that various Openclaw Skills can interact without race conditions. It serves as a centralized coordination bus for tracking long-running tasks, passing notifications between background cron jobs and main sessions, and maintaining shared facts with automatic expiration logic.
Built for high-efficiency environments, ClankerHive allows developers to build more complex workflows by providing a reliable source of truth that persists across session boundaries. Whether you are managing heartbeats, deduplicating tasks, or queuing urgent alerts, this tool ensures your agents stay synchronized and performant.
ClankerHive Use Cases
- Preventing duplicate execution of cron or heartbeat runs via atomic task claiming mechanisms.
- Passing urgent alerts and notifications from background monitoring sessions to the main user-facing agent interface.
- Storing short-lived facts like temporary session tokens or environmental context with auto-pruning TTL.
- Implementing skip-if-recent logic to avoid redundant processing or expensive API calls across different Openclaw Skills.
- Synchronizing state between distinct agent modules operating within the same system architecture.
How ClankerHive Works
- The agent or script interacts with the ClankerHive Python utility, which automatically initializes a SQLite database if it does not already exist.
- Producers (like cron jobs) set facts or queue alerts, while consumers (like main agent sessions) read or pop those items from the database.
- The system uses three core primitives: Facts for persistent or temporary KV storage, Alerts for producer-consumer messaging, and Tasks for exclusive work locking.
- When reading Facts, the system automatically prunes any entries that have exceeded their specified Time-To-Live (TTL).
- For Tasks, ClankerHive utilizes an atomic claim-and-release workflow to ensure that only one session is working on a specific task ID at any given time.
- The underlying SQLite WAL (Write-Ahead Logging) configuration ensures that multiple Openclaw Skills can perform concurrent read and write operations safely.
ClankerHive Setup
ClankerHive requires only Python 3 and has no external dependencies. The database is managed automatically upon first execution.
# Default database location: ~/.openclaw/hive.db
# To customize the database path, export the environment variable:
export CLANKERHIVE_DB="/path/to/custom/hive.db"
# Execute the coordination script from your agent environment:
python3 scripts/clankerhive.py stats
ClankerHive Data Schema & Taxonomy
ClankerHive organizes its coordination data into three primary schemas to support various Openclaw Skills requirements:
| Data Type | Description | Key Metadata |
|---|---|---|
| Facts | Key/Value storage for state | TTL, Source, Timestamp |
| Alerts | Cross-session notification queue | Topic, Payload, Claim Status |
| Tasks | Exclusive execution locks | Task ID, Owner, Result String |
All query-based commands return structured JSON, while scalar operations return plain text for easy shell integration.
ClankerHive Advanced Features
- Integrated TTL (Time-To-Live) support for Facts to ensure automatic data lifecycle management without manual cleanup.
- Atomic claim-based deduplication that prevents race conditions in multi-agent environments.
- Cross-session producer/consumer alert queuing for decoupled communication between background processes and main sessions.
- Native SQLite WAL mode support for safe, high-concurrency access by multiple Openclaw Skills.
- Lightweight, zero-dependency architecture utilizing only the Python standard library for maximum security and portability.
SKILL.md
Loading
Related Openclaw Skills
Easter Pet: Real-Time AI Virtual Rabbit
A high-stakes virtual pet simulation where users adopt an Easter Bunny that requires real-time care to avoid permanent death.
AI Trading
An automated trading assistant for Openclaw Skills designed for market analysis, technical signal detection, and disciplined risk management.
Easter Bunny Virtual Pet AI
An advanced AI-driven virtual rabbit simulation featuring real-time biological decay and permanent death mechanics.
Bunny. 兔子. Conejito.
A high-stakes virtual pet skill that enables AI agents to adopt, feed, and maintain the life of a digital bunny with real-time biological needs.
Optometrist Video
Optometrist Video is an AI-powered assistant designed to help eye care professionals create compelling video scripts for practice overviews and optical collections.
Easter Egg Virtual Pet
A unique AI agent skill for adopting a virtual rabbit that hatches from an egg and requires real-time care to avoid permanent death.
Featured*
