A pre-install security gate that scans Openclaw Skills for prompt injections, malware, and hardcoded secrets before they touch your system.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install skill-guard-snyk-agent-scan
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 skill-guard-snyk-agent-scan 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 skill-guard?
skill-guard serves as the definitive security layer for users interacting with the ClawHub ecosystem. While traditional repositories often rely on basic signature-based antivirus, this tool performs deep content analysis on Openclaw Skills to identify AI-specific threats that standard scanners miss. By wrapping the installation process in a secure pre-flight check, it ensures that your AI agents remain protected from malicious instructions and unauthorized data exfiltration attempts.
The utility leverages the power of Snyk Agent Scan to analyze the actual logic and natural language instructions within a skill. This approach allows developers to maintain a high level of trust in their automation workflows, knowing that every one of their Openclaw Skills has been vetted for toxic flows and hidden prompt injections that could compromise their machine state or sensitive data.
skill-guard Use Cases
- Scanning new Openclaw Skills from ClawHub before finalizing the installation to prevent system compromise.
- Auditing skills for hardcoded secrets or API tokens that might have been accidentally left in the code.
- Detecting hidden prompt injection attacks designed to hijack agent behavior or exfiltrate conversation history.
- Implementing a secure CI/CD pipeline for deploying trusted Openclaw Skills across a developer team.
How skill-guard Works
- The safe-install script is executed with the target skill slug as an argument.
- The utility downloads the requested Openclaw Skills into a temporary staging directory in /tmp/ to isolate them from the primary workspace.
- Snyk Agent Scan is triggered to perform a comprehensive security audit of the skill's content.
- If the scan returns a clean result, the skill is automatically moved to the user's permanent skills directory.
- If vulnerabilities are found, the installation is blocked, the skill is quarantined, and a detailed threat report is generated.
skill-guard Setup
To begin securing your Openclaw Skills, ensure you have the necessary dependencies installed and your Snyk authentication configured:
npm i -g clawhub
curl -LsSf https://astral.sh/uv/install.sh | sh
export SNYK_TOKEN=your_snyk_api_token
Run the secure installation command:
./scripts/safe-install.sh <skill-slug>
skill-guard Data Schema & Taxonomy
The skill-guard utility manages data through a staging lifecycle and clear exit codes to communicate security status. It organizes scan results into specific error and warning categories for granular review.
| Component | Details |
|---|---|
| Staging Directory | /tmp/skill-guard-staging/skills/ |
| Exit Code 0 | Success - Clean install |
| Exit Code 2 | Threat detected - Installation blocked |
| Exit Code 3 | Configuration Error - Missing SNYK_TOKEN |
| Threat IDs | E004 (Injection), E006 (Malware), W007 (Secrets) |
skill-guard Advanced Features
- Client-side pre-flight checks that offer superior protection compared to server-side signature scanning for Openclaw Skills.
- Integrated quarantine system allowing developers to manually inspect suspicious code before deciding to proceed.
- Support for version-specific scanning to ensure that updates to your Openclaw Skills haven't introduced new security regressions.
- Automated detection of AI-specific threat vectors including toxic instruction chaining and machine state compromise attempts.
SKILL.md
Loading
Related Openclaw Skills
Luma Events Enhanced
Automate the discovery and tracking of tech events and meetups globally from Luma.
Skill Vetter
A security-first vetting protocol designed to detect malicious code and risk levels in AI agent skills before installation.
gog: Google Workspace CLI Skill
A comprehensive command-line interface for managing Google Workspace services including Gmail, Calendar, Drive, and Sheets.
WebUntis Timetable Integration
A specialized tool for AI agents to retrieve and summarize school timetables and substitution plans directly from WebUntis.
Masumi Network Warranty Vault
A decentralized warranty verification skill that utilizes the Cardano blockchain to create immutable audit trails for product receipts and agent services.
UBTRIPPIN
A comprehensive travel and event management tool that transforms booking confirmations into structured itineraries for AI agents.
Featured*
