ClawdScan is a diagnostic utility designed to maintain the health of AI agent environments by identifying session bloat and reclaiming storage.
The fastest way to install a skill directly from the registry.
npx clawhub@latest install clawdscan
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 clawdscan using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Get the raw skill files in a ZIP archive.
ClawdScan serves as a vital maintenance utility for developers working with Openclaw Skills. It acts as a zero-dependency Python CLI that deep-scans session JSONL files to pinpoint performance bottlenecks, such as zombie sessions that were never utilized or stale files consuming unnecessary disk space.
By providing granular visibility into how AI models and tools are used, it ensures that your agentic workflows remain lean and efficient. This skill is essential for anyone running long-term AI deployments where session data can quickly accumulate and impact system performance. It offers a suite of commands to inspect individual sessions, track historical trends, and automate the cleanup of problematic data points within the Openclaw Skills ecosystem.
To install ClawdScan as part of your Openclaw Skills library, use the built-in package manager:
clawdbot skill install clawdscan
Alternatively, you can run it as a standalone Python script:
chmod +x clawdscan.py
./clawdscan.py --help
You may also configure environment variables like CLAWDBOT_DIR to point the scanner to a custom directory.
ClawdScan analyzes session data and can export reports in JSON format. It organizes data based on the following taxonomy:
| Data Type | Description | Location |
|---|---|---|
| Session Logs | JSONL files containing message history | agents/*/sessions/ |
| Meta Reports | Summary of health scans and issues | output.json |
| History | Historical trends and growth metrics | CLI output/JSON |
| Metadata | Timestamps, tool IDs, and model names | Extracted from JSONL |
The tool specifically tracks thresholds for bloat (sessions over 1MB or 300 messages), zombies (created but unused for 48 hours), and staleness (inactive for 7+ days).
Loading
A comprehensive toolkit to deploy, optimize, and manage a professional AI development team within Claude Code.

A comprehensive MCP server providing 34 tools for AI agents to interact with Twitter/X through local residential IPs and micropayments.

A headless market intelligence engine for detecting underserved app store niches and generating data-driven MVP specifications.

A high-precision OCR tool that redacts sensitive values from scanned PDF contracts while leaving field labels intact.

An advanced security blacklist system designed to protect AI agents from malicious commands, scams, and prompt injection attacks.

A local, privacy-first CLI tool for Openclaw Skills that tracks LLM token consumption and provides real-time cost analysis across multiple providers.








































