Prompt Guard 作者: Gopikrishnan
A client-side privacy extension that automatically detects and masks sensitive information before you send it to AI chatbots like ChatGPT, Gemini, and Claude.
扫码在 Android 版 Firefox 中打开此扩展
扩展元数据
关于此扩展
Prompt Guard - Secure Your AI Conversations
Short Description:
A client-side privacy extension that automatically detects and masks sensitive information before you send it to AI chatbots like ChatGPT, Gemini, and Claude.
Detailed Description:
Prompt Guard is an essential privacy tool designed to protect your sensitive data (like PII, API keys, and credentials) from accidentally being shared with Large Language Models (LLMs) and AI chatbots.
As AI tools become a daily utility, it is easy to accidentally paste sensitive code, internal company documents, or personal information into a chat prompt. Prompt Guard runs seamlessly in your browser to monitor your inputs locally and prevent data leaks before they happen.
Key Features:
• Real-time Detection: Instantly scans your text inputs on supported AI domains (such as chatgpt.com, gemini.google.com, claude.ai) as you type or paste.
• Automatic Data Masking: When a sensitive pattern is detected, it automatically masks the text (e.g., ****) so it cannot be read by the AI, while allowing the rest of your prompt to remain intact.
• Customizable Rules Engine: Comes pre-loaded with regex rules to detect common secrets (Credit Cards, Social Security Numbers, API Keys, Passwords). You can easily add, edit, or remove your own custom regex patterns in the Options page.
• Severity Levels: Rules are categorized by severity (CRITICAL, HIGH, MEDIUM, LOW) so you can prioritize alerts.
• On-page Notifications: Injects a lightweight, non-intrusive warning notification directly on the page when sensitive data is caught.
• Dashboard & Analytics: The extension popup provides a quick summary of recent alerts, matches, and which domains triggered them.
Privacy by Design:
Prompt Guard is built with a strict zero-data-collection policy:
1. 100% Local Processing: All regex scanning and data masking happens entirely within your browser on your own device.
2. No Telemetry: We do not track your browsing history, we do not log your chat prompts, and we do not collect usage analytics.
3. No External Servers: The extension does not make any external network requests. Your data never leaves your computer.
How to Use:
1. Install the extension.
2. Navigate to your favorite AI chatbot (e.g., ChatGPT).
3. Try typing or pasting sensitive data (like a fake credit card number or an API key).
4. Watch Prompt Guard automatically alert you and mask the sensitive string before you hit send!
5. Click the Prompt Guard icon in your toolbar to view recent detections or click the "Settings" gear to add your own custom rules.
Short Description:
A client-side privacy extension that automatically detects and masks sensitive information before you send it to AI chatbots like ChatGPT, Gemini, and Claude.
Detailed Description:
Prompt Guard is an essential privacy tool designed to protect your sensitive data (like PII, API keys, and credentials) from accidentally being shared with Large Language Models (LLMs) and AI chatbots.
As AI tools become a daily utility, it is easy to accidentally paste sensitive code, internal company documents, or personal information into a chat prompt. Prompt Guard runs seamlessly in your browser to monitor your inputs locally and prevent data leaks before they happen.
Key Features:
• Real-time Detection: Instantly scans your text inputs on supported AI domains (such as chatgpt.com, gemini.google.com, claude.ai) as you type or paste.
• Automatic Data Masking: When a sensitive pattern is detected, it automatically masks the text (e.g., ****) so it cannot be read by the AI, while allowing the rest of your prompt to remain intact.
• Customizable Rules Engine: Comes pre-loaded with regex rules to detect common secrets (Credit Cards, Social Security Numbers, API Keys, Passwords). You can easily add, edit, or remove your own custom regex patterns in the Options page.
• Severity Levels: Rules are categorized by severity (CRITICAL, HIGH, MEDIUM, LOW) so you can prioritize alerts.
• On-page Notifications: Injects a lightweight, non-intrusive warning notification directly on the page when sensitive data is caught.
• Dashboard & Analytics: The extension popup provides a quick summary of recent alerts, matches, and which domains triggered them.
Privacy by Design:
Prompt Guard is built with a strict zero-data-collection policy:
1. 100% Local Processing: All regex scanning and data masking happens entirely within your browser on your own device.
2. No Telemetry: We do not track your browsing history, we do not log your chat prompts, and we do not collect usage analytics.
3. No External Servers: The extension does not make any external network requests. Your data never leaves your computer.
How to Use:
1. Install the extension.
2. Navigate to your favorite AI chatbot (e.g., ChatGPT).
3. Try typing or pasting sensitive data (like a fake credit card number or an API key).
4. Watch Prompt Guard automatically alert you and mask the sensitive string before you hit send!
5. Click the Prompt Guard icon in your toolbar to view recent detections or click the "Settings" gear to add your own custom rules.
评分 0(1 位用户)
权限与数据
必要权限:
- 获取浏览器标签页
- 访问您用于 chatgpt.com 域名的网站的数据
- 访问您用于 claude.ai 域名的网站的数据
- 访问您用于 poe.com 域名的网站的数据
- 访问您在 chatgpt.com 的数据
- 访问您在 chat.openai.com 的数据
- 访问您在 claude.ai 的数据
- 访问您在 gemini.google.com 的数据
- 访问您在 copilot.microsoft.com 的数据
- 访问您在 perplexity.ai 的数据
- 访问您在 www.perplexity.ai 的数据
- 访问您在 poe.com 的数据
可选权限:
- 访问您用于 chatgpt.com 域名的网站的数据
- 访问您用于 claude.ai 域名的网站的数据
- 访问您用于 poe.com 域名的网站的数据
- 访问您在 chatgpt.com 的数据
- 访问您在 chat.openai.com 的数据
- 访问您在 claude.ai 的数据
- 访问您在 gemini.google.com 的数据
- 访问您在 copilot.microsoft.com 的数据
- 访问您在 perplexity.ai 的数据
- 访问您在 www.perplexity.ai 的数据
- 访问您在 poe.com 的数据
收集的数据:
- 开发者称此扩展无需收集数据。
更多信息