A platform-aware, zero-dependency system cleanup utility designed to safely reclaim disk space for AI agents.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install claw-janitor
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 claw-janitor 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 Claw Janitor?
Claw Janitor is a specialized utility within the Openclaw Skills ecosystem designed for intelligent system maintenance. It serves as a safe, platform-aware janitorial service for your development environment, ensuring that disk space constraints never interrupt your AI agent's productivity. By targeting temporary files, package manager caches, and Docker remnants, it keeps your workspace lean without risking critical data.
The skill is built with a safety-first philosophy, incorporating built-in protections for symlinks, mounts, and sensitive AI caches. Whether you are running a routine cleanup or an aggressive deep-cleaning session, Claw Janitor provides the transparency needed through dry-run modes and detailed JSON reporting, making it an essential addition to any suite of Openclaw Skills.
Claw Janitor Use Cases
- Reclaiming disk space when storage is low during intensive AI development.
- Automating the removal of stale Docker images and unused package manager caches.
- Generating audit reports of system clutter before performing manual deletions.
- Performing safe, non-destructive cleanups of temporary system files across different operating systems.
How Claw Janitor Works
- The skill identifies the host operating system to apply platform-specific cleanup logic.
- It scans defined target groups such as ai-scan, packages, docker, and system files.
- It cross-references found items against a safety blacklist to protect critical system paths and AI-cache files.
- If dry-run is enabled, it outputs a preview of actions; otherwise, it executes the cleanup.
- A summary report is generated, optionally saved to a JSON file for audit trails.
Claw Janitor Setup
To integrate this tool into your Openclaw Skills workspace, ensure the script is located in your skills directory and execute it via Node.js:
# Basic installation check
node /root/.openclaw/workspace/skills/claw-janitor/janitor.js --help
# Run a safe dry-run to verify paths
node /root/.openclaw/workspace/skills/claw-janitor/janitor.js --dry-run
Claw Janitor Data Schema & Taxonomy
Claw Janitor organizes its operations into specific functional groups and can output structured data for integration with other tools:
| Group | Description |
|---|---|
| ai-scan | Analyzes and cleans temporary AI-related artifacts (excluding protected caches). |
| packages | Targets package manager caches like npm, pip, or apt. |
| docker | Cleans up dangling images, stopped containers, and unused networks. |
| system | Handles OS-level temporary files and logs. |
When using the --report-file <path> flag, the skill generates a JSON schema containing timestamp, total_space_reclaimed, and a details array of deleted items.
Claw Janitor Advanced Features
- Dry-run mode for risk-free simulation of cleanup tasks.
- Targeted execution using
--onlyor--skipto focus on specific resource groups. - Structured JSON output for multi-agent orchestration and automated monitoring.
- Deep-cleaning mode for aggressive reclamation of system resources.
- Platform-aware logic that automatically adjusts paths for different environments.
SKILL.md
Loading
Related Openclaw Skills
Oura API Complete Export
A professional-grade data extraction tool that automates Oura Ring OAuth2 authentication and exports all available v2 user collection datasets to raw JSON.
TeamClaw
A high-performance multi-agent orchestration subsystem featuring the OASIS collaboration engine, automated computer use, and social platform integrations.
Memphis CLI
A local-first brain management CLI for capturing, querying, and synchronizing immutable memory chains across AI agents.
Teamclaw
A high-performance Agent subsystem designed for complex multi-agent orchestration, automated scheduling, and real-time visual monitoring.
Video Translation
An automated workflow to translate video speech and replace the original audio with high-fidelity, synchronized dubbed audio.
Tavily Best Practices
A comprehensive guide and toolset for integrating Tavily's real-time web search and research capabilities into production-ready AI agents.
Featured*
