A deterministic redaction engine that prevents private infrastructure details from leaking into public messaging channels.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install content-scrubber
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 content-scrubber 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 Content Scrubber?
Content Scrubber is a specialized extension designed for users of Openclaw Skills who require high-level privacy and security. It acts as a safety barrier by intercepting outgoing communications and scrubbing sensitive technical data—such as internal IP addresses and hostnames—before they are transmitted to external platforms like Discord, Telegram, or Signal.
Unlike traditional LLM-based redaction which can be unpredictable, this skill utilizes a robust, regex-driven architecture. This ensures that privacy rules are applied with deterministic precision, making it an essential tool for developers and sysadmins who want to share agent outputs without exposing the internal topology of their private networks.
Content Scrubber Use Cases
- Sharing logs or command outputs from Openclaw Skills to public Discord servers without exposing internal server IPs.
- Collaborating in shared channels while hiding specific SSH targets and user credentials.
- Testing AI agent responses in a dryRun mode to audit potential data leaks before they happen.
- Protecting localhost service ports and loopback addresses during live demonstrations or troubleshooting sessions.
How Content Scrubber Works
- The plugin registers as a high-priority message interceptor within the Openclaw environment.
- Every outgoing message is scanned against a comprehensive library of regex patterns targeting RFC 1918 IPs, loopback addresses, and SSH syntax.
- Detected sensitive strings are replaced with standardized placeholders like [redacted-ip] or [redacted-target].
- The sanitized message is passed forward to the messaging provider (e.g., Telegram API) for final delivery.
- If enabled, the dryRun mode logs the detected patterns to the console without altering the message content.
Content Scrubber Setup
- Deploy the plugin files to your local extension directory:
mkdir -p ~/.openclaw/extensions/content-scrubber
# Copy index.ts, openclaw.plugin.json, and package.json to this folder
- Update your global configuration file (
openclaw.json) to activate the skill:
{
"plugins": {
"entries": {
"content-scrubber": {
"enabled": true,
"config": {
"dryRun": false,
"allowedRecipients": []
}
}
}
}
}
- Restart the Openclaw service to apply the new security layers.
Content Scrubber Data Schema & Taxonomy
The Content Scrubber manages its lifecycle through the following configuration schema and metadata:
| Option | Type | Description |
|---|---|---|
dryRun |
boolean | If true, identifies sensitive data in logs but does not modify the outgoing message. |
allowedRecipients |
array | A whitelist of Chat IDs where the scrubbing logic is bypassed. |
placeholders |
internal | Mapping of regex hits to labels like [redacted-service] or [redacted-ip]. |
All transformations are performed in-memory during the interception phase, ensuring no unredacted logs are persisted to disk by this specific skill.
Content Scrubber Advanced Features
- Deterministic Regex Engine: Guaranteed protection against edge cases that often bypass probabilistic LLM scrubbers.
- Recipient Whitelisting: Granular control to skip redaction for trusted private DMs while maintaining strict rules for group channels.
- Multi-Protocol Support: Seamlessly works across all messaging surfaces supported by Openclaw Skills, including Signal and Telegram.
- Audit Logging: Use dryRun to perform security audits of your AI agent's communication habits without interrupting the user experience.
SKILL.md
Loading
Related Openclaw Skills
China Image Generator (china-image-gen)
A specialized image generation skill that leverages the SiliconFlow API to provide high-performance AI art creation within China without needing a VPN.
Ubuntu Encyclopedia
A documentation-first workflow agent for Ubuntu system administration, maintenance, and expert-level troubleshooting.
Skill Usage Analyzer
A powerful utility designed to parse SKILL.md files and generate actionable, human-friendly usage guides for any AI agent skill.
TikTok Live Reply Scripts
A specialized AI skill designed to generate real-time, high-conversion objection handling scripts for TikTok Shop live selling sessions.
Research Agent
A sophisticated AI research specialist that transforms vague inquiries into comprehensive, source-backed analysis reports and comparison matrices.
Income Tracker
A comprehensive financial management tool for freelancers and creators to track multi-source income with automated currency conversion and visual analytics.
Featured*
