PromptShield - AI Prompt Security
kminnovations.dev · Privacy & Security
Media
5 items
Description
Block API keys, secrets, tokens, and sensitive data before sending prompts in supported AI chat sites. 100% local. PromptShield helps prevent accidental secret sharing in AI prompt boxes. It checks prompt text locally in your browser before a message is sent. When risky content is found, PromptShield can warn, block, or redact sensitive values so you can review the prompt safely. It can detect API keys, access tokens, private keys, JWTs, bearer tokens, .env-style secrets, email addresses, phone numbers, internal URLs, staging links, localhost links, custom domains, and custom watchwords. Key features: - Local detection in your browser - No backend service - No prompt uploads - No prompt history stored - Redaction before review - Normal and strict protection modes - Per-site pause controls - Custom domains and watchwords - Local scan, warning, block, and redaction counters PromptShield uses chrome.storage.local only for settings and basic counters. It does not use external AI APIs, analytics, or telemetry.