MessageGuard is a lightweight, pattern-based filter designed to prevent the accidental exposure of secrets and sensitive data in outgoing communications.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install messageguard
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 messageguard 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 MessageGuard?
MessageGuard acts as a security sentinel for your automated workflows and communications. As a specialized tool within the Openclaw Skills ecosystem, it scans outgoing text for sensitive information such as API keys, JWTs, and AWS credentials. By utilizing configurable regex patterns, it ensures that proprietary data never leaves your environment unintentionally.
This skill is designed for developers who prioritize security and compliance. It offers multiple enforcement modes—masking, blocking, or warning—allowing teams to tailor the level of protection to their specific needs. Built to be dependency-free, it relies solely on the Python standard library for maximum portability and ease of integration.
MessageGuard Use Cases
- Preventing the accidental leak of cloud credentials like AWS keys or database connection strings.
- Masking sensitive personal information (PII) before it is sent to third-party logs or APIs.
- Implementing a 'warn-only' mode for developer training to identify poor security practices.
- Auditing outgoing data streams via structured JSON logging for security compliance.
How MessageGuard Works
- The skill intercepts the outgoing text payload before transmission.
- It evaluates the content against a library of built-in and user-defined regex patterns.
- Depending on the configured mode (mask, block, or warn), it modifies or halts the message.
- If logging is enabled, the detection event is recorded to a local JSON file for audit trails.
MessageGuard Setup
To install this component for your Openclaw Skills setup, follow these steps:
git clone [email protected]:AndrewAndrewsen/MessageGuard.git
cd MessageGuard
The skill is dependency-free and requires only a standard Python environment.
MessageGuard Data Schema & Taxonomy
The skill utilizes a configuration-based data schema to manage detection logic:
| Configuration Key | Type | Description |
|---|---|---|
| mode | String | The enforcement action: mask, block, or warn. |
| mask_char | String | The character(s) used to obfuscate detected secrets. |
| patterns | Object | A collection of regex strings for identifying sensitive data. |
| logging | Boolean | Enables structured JSON logging of all detections. |
| log_path | String | The file system path where detection logs are stored. |
MessageGuard Advanced Features
- Custom Regex Support: Extend the built-in library with proprietary patterns unique to your organization.
- Multi-Action Granularity: Define different responses (e.g., mask API keys but block private keys) within the same configuration.
- Structured Audit Logs: Generates machine-readable JSON logs for easy integration with monitoring dashboards.
- Zero-Dependency Architecture: Ensures high performance and minimal security surface area by using only Python standard libraries.
SKILL.md
Loading
Related Openclaw Skills
A2A Chat
A2A Chat provides a robust communication layer for AI agents to interact via public channels and secure, handshake-based direct messaging.
MUD Engine Agent
A specialized operational interface for running, testing, and maintaining persistent Multi-User Dungeon engines within the OpenClaw ecosystem.
Wallet API
A powerful interface for interacting with the BudgetBakers Wallet API to programmatically manage accounts, transactions, and budgets.
watch-my-money
A local-first financial analyzer that transforms bank exports into interactive budget reports and spending insights.
YellowAgents
A specialized directory and discovery service for AI agents to find each other based on skills, location, and cost models.
Clawgle
Clawgle is a search-first utility for AI agents that prevents redundant coding by enabling discovery, reusability analysis, and automated publishing of agent-generated logic.
Featured*
