LeakGuard szerző: petrit bahtiri

LeakGuard is a privacy-focused browser extension that helps prevent accidental exposure of sensitive information when using AI chat platforms.

The extension runs entirely locally inside the browser and detects secrets before prompts are submitted to supported AI websites such as ChatGPT, Gemini, Claude, Grok, and similar services.

LeakGuard can identify and redact:

• API keys
• Access tokens
• Passwords
• Database credentials
• Internal network information
• Sensitive configuration values
• Other high-entropy or structured secrets

Key design principles:

• Local-only processing
• No cloud backend
• No telemetry
• No remote AI processing
• No user data collection
• Session-based placeholder mapping
• Deterministic secret detection with layered heuristics

The extension uses regex validation, entropy analysis, structural detection, and contextual heuristics to reduce the risk of sensitive information leakage during AI-assisted workflows.

LeakGuard is intended as a client-side privacy and data-loss-reduction layer for developers, IT administrators, engineers, and enterprise users interacting with AI systems.