A durable, SQLite-backed orchestration runtime for scheduling multi-step shell commands and AI agent workflows with automated retries and manual approvals.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install durable-scheduler
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 durable-scheduler using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
Openclaw Scheduler is an advanced, enterprise-grade orchestration runtime designed for developers and operators who need robust, audit-logged workflow scheduling. It functions as a standalone daemon (ai.openclaw.scheduler) backed by SQLite, ensuring that tasks survive gateway restarts or system failures. By decoupling complex automations from basic built-in schedulers, it enables seamless execution of both standard shell commands and intelligent agent tasks.
As a crucial addition to the ecosystem of Openclaw Skills, this scheduler excels at executing multi-step pipelines. It safely manages dependencies, handles retries gracefully, supports manual human-in-the-loop approval gates, and maintains a complete, persistent execution history. It is the ideal tool for production-grade automation that demands absolute visibility, reliability, and granular control.
ai.openclaw.scheduler) using Node.js 22+ with a lightweight SQLite database to maintain state.waitForGateway loop, holding the job safely in the queue until communication is restored.Follow these steps to initialize and run the scheduler on your system:
Install the package globally using npm:
npm install -g openclaw-scheduler
Initialize the database and configuration files, creating the SQLite database and launch agent configuration:
openclaw-scheduler init
Bootstrap the service using launchd (macOS/Unix environments):
launchctl bootstrap gui/$(id -u) ~/Library/LaunchAgents/ai.openclaw.scheduler.plist
Verify that the scheduler is up, running, and communicating with the database:
openclaw-scheduler jobs list
Define a scheduled agent job:
openclaw-scheduler jobs add \
--name "weekly digest" \
--cron "0 9 * * 1" \
--kind agent \
--to "@channel-id" \
--prompt "Summarize last week's incidents and post to #digest"
The scheduling runtime stores state, job definitions, and run logs locally. Here is how its data architecture is structured:
| Location | File / Resource | Description |
|---|---|---|
~/.openclaw/scheduler/ |
Directory | Root folder containing all configuration, job, and execution data. |
~/.openclaw/scheduler/scheduler.db |
SQLite Database | Stores configurations, workflow definitions, states, retries, and step transition records. |
~/Library/LaunchAgents/ai.openclaw.scheduler.plist |
launchd Service Plist | Service daemon configuration mapping to Node.js 22+ runtime. |
Every executed job preserves standard operational details including run duration, step indices, transition timestamps, and explicit stdout/stderr outputs to enable developer debugging via the command-line interface.
shell:check -> agent:diagnose -> approval -> shell:remediate.waitForGateway(timeoutMs, intervalMs) logic, allowing scheduled jobs to survive gateway restarts without failing silently.--timeout 30m) and custom triggers for approved or rejected transitions.Loading
A resilient framework designed to safely restart, update, or reload the OpenClaw gateway while maintaining conversation continuity across various messaging channels.

An AI-powered design and development framework to systematically restructure vocational education courses based on professional competency mapping.

An automated daily news aggregator that leverages Openclaw Skills to collect, cross-verify, and report global events across eight domains.

A respectful relationship communication helper that interprets emotional subtext and drafts constructive, non-defensive replies.

An advanced AI agent persona and quantitative-heuristic reasoning engine that replicates the practical A-share trading framework of private fund manager Xu Zhexun.

An automated review skill for auditing Remix v2 codebases to prevent breaking meta tag changes, cookie security gaps, improper auth gates, and missing CSRF protections.
