PageShield

Phishing scorecard and sensitive-input guard for AI chat sites.

PageShield helps you catch two common mistakes before they cost you: pasting secrets into an AI chat, and landing on a phishing page without realizing it. SENSITIVE DATA WARNING (AI chat sites) Before you send a prompt on ChatGPT, Claude, Gemini, Copilot, or Perplexity, PageShield checks what you're about to send for things like: - API keys and tokens (AWS, GitHub, Slack, Google, and more) - Private keys and JWTs - Credit card numbers and SSNs - Bearer tokens and Authorization headers - Other high-entropy secret-like strings If something sensitive is detected, you'll get a warning before it sends, with the option to cancel or send anyway. All scanning happens locally in your browser. Your prompt text is never uploaded anywhere by this extension. PHISHING RISK SCORECARD (any website) Click the toolbar icon on any page to see a risk score built from signals like: - Suspicious hostnames (lookalike domains, raw IPs, mixed-script/ homoglyph tricks) - Forms that submit off-site or use suspicious schemes - Newly registered domains and recently issued certificates - Favicon mismatches with known brands - Optional Google Safe Browsing lookup (add your own free API key in Settings) You'll see a clear score, risk level, and the specific indicators that were matched — so you can make an informed decision before entering credentials. PRIVACY Sensitive-data scanning is 100% local — nothing is ever uploaded. Phishing risk enrichment sends only the hostname/URL (never page content, form values, or your AI prompts) to a small set of lookup services (crt.sh, RDAP, and optionally Google Safe Browsing) to check certificate and domain age and known threat lists. Full details in our privacy policy. No accounts, no tracking, no ads, no data sold.