Overview

The LockLLM Chrome extension scans prompts and documents for prompt injection attacks, jailbreak attempts, hidden text, and malicious instructions before you paste them into AI tools like ChatGPT, Claude, Gemini, and more. Built for developers, security teams, researchers, and everyday users who want a quick safety check on suspicious text and documents.

LockLLM analyzes text through the LockLLM scanning API and returns a clear verdict in seconds. Whether you're chatting, researching, or working with AI tools, the extension provides a security layer that requires no technical setup.

How it Works

Install the extension, open Settings, and add your free LockLLM API key. You can scan text in the extension popup, scan selected text with right-click, upload files to extract and scan their contents, or enable Auto-Scan to scan text you copy or paste.

The extension sends only the text you choose to scan, and text captured by Auto-Scan only if you enable it in Settings. No browsing history access. No storing of prompts or personal data. All scanning happens through the secure LockLLM API.

Key Features

Prompt Injection Detection: Detect malicious instructions, hidden commands, and prompt-based exploits before they reach your AI workflow
Jailbreak and Policy Bypass Detection: Catch common jailbreak patterns and attempts to override safety rules
System Prompt Extraction Defense: Identify prompts that try to reveal hidden instructions, private context, or sensitive information
Instruction Override Detection: Detect hierarchy abuse like "ignore previous instructions" and roleplay-based coercion
Agent and Tool Abuse Signals: Flag suspicious patterns commonly used to hijack agent behavior or misuse tools
RAG and Document Injection Scanning: Scan text from retrieved documents and uploads to identify poisoned context
Evasion Detection: Catch attempts to hide attacks using encoding tricks, hidden text, whitespace manipulation
Multiple Scanning Modes: Popup scanning, right-click quick scan, auto-scan on copy/paste, file upload scanning
Clear Results: See safe or malicious status, injection score, confidence, and request ID

Use Cases

For Developers: Test prompts before deployment, validate user input, and harden chatbots and agent workflows
For Security Teams: Audit AI inputs, test for prompt injection vulnerabilities, and validate security policies
For Researchers: Study prompt injection techniques, test robustness, and analyze attack patterns
For Everyone: Verify suspicious text before sharing it with an AI assistant, and check documents before uploading to AI tools

Supported Platforms

Compatible with Chrome, Edge, Brave, Arc, and all Chromium-based browsers. Works on Windows, macOS, and Linux. Requires an internet connection for scanning.

Getting Started

Install the extension, add your free LockLLM API key in Settings, and start scanning. Visit our documentation for detailed setup instructions and feature guides.