A lightweight security utility designed to detect and block malicious commands or prompt injections within AI agent data streams.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install input-validator
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 input-validator 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 Input Validator?
Input Validator is a specialized security component for AI coding agents, providing a robust layer of protection against malicious inputs. By analyzing text from web pages, uploaded files, and API responses, it identifies dangerous patterns like system-level commands, unauthorized downloads, or unauthorized system access without disrupting the normal developer experience.
As a core part of the Openclaw Skills ecosystem, this tool ensures that automated workflows remain safe from both external threats and internal prompt injection attempts. It is designed to be gentle and non-intrusive, only flagging content that poses a genuine risk while maintaining high-speed performance for real-time applications.
Input Validator Use Cases
- Validating scraped content after web fetching to prevent executing hidden malicious scripts or shell commands.
- Scanning user-uploaded files for system-damaging patterns like recursive deletions or permission escalations.
- Monitoring RSS feeds and external API responses for suspicious strings or unexpected administrative instructions.
- Neutralizing prompt injection and jailbreak attempts by identifying phrases designed to bypass safety filters.
How Input Validator Works
- The skill receives raw input text, a file path, or an API response for security analysis.
- It performs a high-speed regex scan against ten categories of dangerous content, including shell injections and miners.
- If a critical threat is identified, the skill immediately flags the content as unsafe and provides a block list.
- The system then evaluates the input for suspicious patterns related to jailbreaking and prompt manipulation.
- A structured result is returned to the agent, allowing for conditional logic based on the safety status of the data.
Input Validator Setup
To get started with this component in the Openclaw Skills environment, ensure Python 3 is available on your system. You can run the validator directly from the command line:
# Validate a simple string
python3 scripts/input-validator.py "Check this content for safety"
# Validate a specific local file
python3 scripts/input-validator.py --file dangerous-script.txt
Input Validator Data Schema & Taxonomy
The skill returns a standardized JSON object that can be easily parsed by other agents or scripts.
| Field | Type | Description |
|---|---|---|
| safe | Boolean | Final safety determination (False if dangerous patterns exist). |
| warnings | Array | List of suspicious patterns detected, such as jailbreak attempts. |
| dangerous | Array | List of critical security threats detected, such as 'rm -rf' commands. |
Input Validator Advanced Features
- Customizable pattern matching allows developers to add new regex rules to the dangerous or suspicious categories.
- Multi-agent integration support enables the validator to act as a middleware for other Openclaw Skills.
- Strict mode toggle for higher-security environments that require more aggressive filtering.
- Lightweight performance profile with sub-50ms execution time and less than 5MB of memory usage.
SKILL.md
Loading
Related Openclaw Skills
GitHub Operations (github-ops)
A fully autonomous skill for managing GitHub repositories, pushing code, and handling releases without user intervention.
Agent Optimizer
A lightweight performance optimization framework for AI agents based on trajectory analysis and reward feedback.
ClawHub - YouTube Video Clipper & Downloader
A specialized utility for extracting specific video segments and audio clips from YouTube using precise timestamp ranges.
Aleph Cloud Self-Deployment
A sophisticated skill enabling AI agents to autonomously provision, configure, and replicate themselves on the Aleph Cloud decentralized network.
Quantitative Trading Signals
An automated quantitative analysis engine that generates multi-indicator trading signals for global equity and cryptocurrency markets.
Stock & Crypto Real-time Monitor (HK/US/Crypto)
A technical analysis tool for real-time monitoring of Hong Kong stocks, US stocks, and cryptocurrencies using Yahoo Finance.
METADATA
Featured*
