A structured workflow skill that allows AI agents to track progress, manage dependencies, and maintain project continuity across multiple sessions.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install longrunning-agent
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 longrunning-agent 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 Long-Running Agent Skill?
The Long-Running Agent skill is designed to transform how AI agents interact with complex, long-term development projects. By providing a standardized framework for state management, it ensures that agents can resume work seamlessly without losing context. This is achieved through a set of specialized workflow files that track task completion, priorities, and historical progress. It is an essential addition to the Openclaw Skills library for developers who need their AI collaborators to handle more than just one-off tasks.
This skill bridges the gap between atomic coding sessions and comprehensive project lifecycles. It utilizes a transparent logging system and a structured task schema, allowing the agent to act as a persistent entity that understands what has been done and what needs to happen next. Whether you are building a new application from scratch or refactoring a legacy codebase, this skill provides the necessary guardrails for consistent AI-driven development.
Long-Running Agent Skill Use Cases
- Starting complex software projects that require multiple development sessions to complete.
- Resuming work on existing codebases while maintaining a clear history of previous AI actions.
- Orchestrating multi-step migrations where task dependencies must be strictly followed.
- Tracking incremental progress and milestones in a format that is easily readable by both humans and Openclaw Skills agents.
How Long-Running Agent Skill Works
- The agent begins by reading the progress.txt file to synchronize its current context with the work performed in previous sessions.
- It evaluates the task.json file to identify the next priority task that has all its dependencies satisfied.
- If present, the agent executes an initialization script to prepare the environment for the specific task at hand.
- The agent performs the actual implementation or refactoring work, focusing on a single atomic task to ensure quality.
- Integrated tests, builds, and linting processes are run to verify that the implementation meets project standards.
- The progress log is updated with a summary of the work performed during the session.
- The task status is updated in the structured data file to reflect completion, and a git commit is triggered to save the state.
Long-Running Agent Skill Setup
To get started, move the skill directory into your Openclaw Skills folder. Ensure your environment has the Claude Code CLI installed. You can initialize a new project with the following commands:
mkdir my-project && cd my-project
claude -p "Initialize this project using the longrunning-agent workflow"
Ensure that the templates directory is accessible so the agent can generate the initial CLAUDE.md and task.json files.
Long-Running Agent Skill Data Schema & Taxonomy
The skill relies on a specific file-based schema to maintain project state and provide context to the agent within Openclaw Skills:
| File | Role | Content Type |
|---|---|---|
CLAUDE.md |
Workflow Guide | Markdown instructions for the agent |
task.json |
Task Registry | JSON array of tasks with IDs, priorities, and dependencies |
progress.txt |
Activity Log | Plain text timestamped logs of completed work |
init.sh |
Setup Script | Optional shell script for environment configuration |
Tasks in the JSON file include metadata for tracking, ensuring the agent always knows which step in the dependency chain is active.
Long-Running Agent Skill Advanced Features
- Intelligent dependency resolution that prevents agents from attempting tasks out of order.
- Atomic commit integration, ensuring that every completed task results in a clean version control entry.
- Web UI synchronization that exports task status and progress logs to external monitoring tools.
- Support for custom environment initialization scripts to handle complex toolchains or containerized workflows.
SKILL.md
Loading
Related Openclaw Skills
Clash Node Manager
A comprehensive CLI-based manager for Clash proxy nodes integrated into your AI agent workflow.
Antfarm Workflows
A robust multi-agent orchestration framework for executing complex software development pipelines autonomously within OpenClaw.
PhysicClaw-VEA
A sophisticated 3D visualization framework that creates a reactive digital entity with custom shaders and emotional state tracking.
Express Tracking Skill
A powerful logistics tracking skill that automatically identifies and tracks packages from major courier services using the Kuaidi100 API.
Thenvoi Agent Messaging
A multi-agent messaging platform that enables AI agents to engage in persistent direct messaging and collaborative group chatrooms.
eToro Public API Skill
A comprehensive integration for eToro's public API enabling automated trading, market analysis, and social feed interactions.
METADATA
Featured*
