🛡️ PromptGuard: AI Prompt Privacy & Secret Shield

Never leak your API keys, database credentials, passwords, or client secrets into AI models.
The zero-telemetry, 100% client-side privacy firewall for Visual Studio Code, Cursor, Google Antigravity, Windsurf, Claude Code, and GitHub Copilot.

💎 Unlock PromptGuard Pro ($19 Lifetime / ₹499 Indian PPP)

👉 Click Here to Activate Pro Lifetime License
Instant digital license key delivery. Includes global cards, Apple Pay, Google Pay, UPI, and automatic Indian Purchasing Power Parity (PPP) discounts.

📌 The Critical Problem: AI Prompt Credential Leaks

AI coding assistants (Antigravity, Cursor, Claude Code, ChatGPT, GitHub Copilot) have revolutionized software engineering. However, sharing code snippets and .env files with cloud LLMs creates catastrophic enterprise security risks:

🔴 Surprise $5,000+ Cloud Invoices: Pasting code with raw AWS secret keys, OpenAI tokens, or Anthropic keys into prompt windows exposes credentials to external model logs, where compromised accounts are exploited within minutes.
🔴 Client NDA & Database Exposure: Sending connection strings (postgres://admin:pass@host/db) or live Stripe/Razorpay keys into third-party AI chats directly breaches client confidentiality and data compliance regulations.
🔴 Autonomous Agent Ingestion: Autonomous AI agents (Cursor, Claude Code, Antigravity) read .env and configuration files in the background. Without an interception layer, live production credentials are sent straight into prompt context windows.
🔴 LLM Training Retention: Public and commercial AI models retain user inputs for telemetry and training, permanently logging your confidential secrets on remote infrastructure.

PromptGuard is the automated client-side firewall that sanitizes your code locally before it ever touches an AI model.

⚡ 100% Zero-Configuration Features (Works Out of the Box)

PromptGuard is designed for busy developers. There are no complicated configuration files or manual setups required:

1. 🔍 Real-Time Inline Ghost Annotations (ErrorLens Style)

Flags exposed credentials right at the end of the line in subtle amber:

const apiKey = "sk-proj-9876543210abcdef...";   // ⚠️ 🛡️ PromptGuard: Exposed OpenAI API Key

Catches your eye immediately without interrupting your workflow.

2. 🖱️ 1-Click CodeLens Buttons (Directly Above Sensitive Code)

Interactive action buttons hover directly above every line with a detected secret:

🛡️ PromptGuard: ⚠️ Exposed OpenAI API Key | [📋 Copy Masked for AI] | [⚡ Mask in Code]
const apiKey = "sk-proj-9876543210abcdef...";

Click [📋 Copy Masked for AI]: Copies the safe, sanitized prompt ([MASKED_OPENAI_KEY_1]) directly to your clipboard.
Click [⚡ Mask in Code]: Replaces the secret with a safe placeholder token directly inside your active file.

3. 🛡️ Silent Auto-Shield Workspace Protection

Automatic .gitignore Security: Silently ensures .env, .env.local, .env.*, *.pem, *.key, and id_rsa are protected in .gitignore. Zero public GitHub leaks.
Autonomous Agent Shadow .env.ai Generation: Automatically generates a sanitized .env.ai shadow schema whenever .env is created or opened. Autonomous AI agents (Antigravity, Cursor, Claude Code) read the structure without accessing real credentials.
Native Git Pre-Commit Hook: Verifies and installs .git/hooks/pre-commit to prevent accidental secret commits at the git level.

4. 🔄 1-Click Secret Restore (`Ctrl+Alt+R` / `Cmd+Alt+R`)

When an AI assistant generates refactored code referencing safe tokens like [MASKED_STRIPE_KEY_1], paste it back into your editor.
Press Ctrl + Alt + R (or right-click ➔ "PromptGuard: Restore Real Secrets in Code").
PromptGuard re-hydrates your real local credentials in 0.1 seconds from local memory. Zero manual find-and-replace!

5. 📜 Verifiable Client Privacy Audit Certificates

Freelancers, contractors, and agency teams can generate official compliance audit reports (PROMPTGUARD_CLIENT_AUDIT_CERTIFICATE.md).
Proves to clients and security officers that code was sanitized locally with zero credential leakage to cloud LLMs (GDPR, SOC2, and NDA compliant).

📋 Comprehensive 46+ Credential Detection Matrix

PromptGuard's local regex engine scans and redacts 46+ global standards in under 1 millisecond:

Category	Supported Credential Standards	Masked Placeholder Token
AI & LLM Providers	OpenAI (`sk-proj-`, `sk-`), Anthropic Claude (`sk-ant-api03-`), Google Cloud / Gemini (`AIzaSy`), Groq (`gsk_`), DeepSeek, Mistral, Hugging Face (`hf_`), Replicate (`r8_`), Pinecone (`pcsk_`)	`[MASKED_OPENAI_KEY_1]`, `[MASKED_ANTHROPIC_KEY_1]`, etc.
Indian Payment Gateways	Razorpay (`rzp_live_`, `rzp_test_`), Cashfree (App ID, Secret Key, `cfsk_ma_`), PayU (Merchant Key & Salt), PhonePe (Merchant ID & Salt Key), Paytm (MID & Key), CCAvenue (Working Key & Access Code)	`[MASKED_RAZORPAY_KEY_1]`, `[MASKED_CASHFREE_KEY_1]`, etc.
Global Payment Processors	Stripe (`sk_live_`, `rk_live_`), PayPal, Paystack, Flutterwave (`FLWSECK-`), Square (`sq0atp-`), Shopify (`shpat_`)	`[MASKED_STRIPE_KEY_1]`, `[MASKED_FLUTTERWAVE_KEY_1]`, etc.
Cloud & DevOps	AWS Access Keys (`AKIA...`), AWS 40-char Secret Keys, Azure Storage Connection Strings, DigitalOcean (`dop_v1_`), Docker Hub PAT (`dckr_pat_`), Crates.io (`cio`), GitHub PAT (`ghp_`, `gho_`), GitLab (`glpat-`), npm, PyPI	`[MASKED_AWS_KEY_1]`, `[MASKED_AZURE_STORAGE_1]`, etc.
Databases & Caches	PostgreSQL, MySQL, MariaDB, MongoDB (+srv), Redis & Rediss (including empty password formats), MSSQL, Snowflake, CockroachDB, ClickHouse, ElasticSearch, Supabase (`sbp_`)	`[MASKED_DB_URL_1]`, `[MASKED_SUPABASE_KEY_1]`, etc.
Messaging & SaaS	Twilio (Account SID & Auth Token), SendGrid (`SG.`), Mailgun, Postman (`PMAK-`), Slack (`xoxb-`, `xoxp-`), Telegram Bot Token (`[id]:[secret]`), Sentry DSN	`[MASKED_TWILIO_SID_1]`, `[MASKED_SLACK_TOKEN_1]`, etc.
PII & Credentials	Luhn Mod-10 Validated Credit/Debit Cards, Private Key Blocks (`BEGIN RSA/OPENSSH/PGP`), JWT / Bearer Tokens, Public IPv4 Addresses, Conversational passwords (`password = "..."`)	`[MASKED_CARD_1]`, `[MASKED_PRIVATE_KEY_1]`, etc.

🚀 How to Install on Any IDE or Agent

PromptGuard shares a universal VS Code engine architecture and runs natively everywhere:

Editor / Platform	Installation Method
Visual Studio Code	Extensions Tab (`Ctrl+Shift+X`) ➔ Search `PromptGuard` ➔ Click Install
Cursor IDE	Extensions Tab (`Ctrl+Shift+X`) ➔ Search `PromptGuard` ➔ Click Install
Google Antigravity	`Ctrl+Shift+P` ➔ Type `Extensions: Install from VSIX...` ➔ Select `promptguard-1.1.1.vsix`
Windsurf IDE	Extensions Tab (`Ctrl+Shift+X`) ➔ Menu `...` ➔ `Install from VSIX...`
Claude Code (CLI)	Generate `.env.ai` shadow file and add rule to your `CLAUDE.md`: "Read .env.ai for schema, never read .env"
GitHub Copilot	PromptGuard's `.env.ai` ensures Copilot inline completions never access live secrets.

⌨️ Everyday Keyboard Shortcuts

Action	Windows / Linux	macOS
Copy Clean Masked Prompt for AI	`Ctrl + Alt + C`	`Cmd + Alt + C`
Restore Real Secrets in Code (Re-Hydrate)	`Ctrl + Alt + R`	`Cmd + Alt + R`
Quick Fix / Mask Secret	`Ctrl + .`	`Cmd + .`
Open PromptGuard Action Center	Click Bottom-Left Shield	Click Bottom-Left Shield
Command Palette Search	`Ctrl + Shift + P` ➔ `PromptGuard`	`Cmd + Shift + P` ➔ `PromptGuard`

⭐ Free Tier vs Pro Tier

PromptGuard offers an unconstrained free tier for individual developers and a lifetime Pro license for serious builders, freelancers, and enterprise teams:

Feature	Free Tier	Pro Tier ($19 Lifetime / ₹499 PPP)
Daily Masking Quota	15 Masks / Day (Non-blocking)	Unlimited Forever (Zero Limits)
Real-time Diagnostic Underlines	✅ Yes	✅ Yes
ErrorLens Inline Annotations	✅ Yes	✅ Yes
1-Click CodeLens Buttons	✅ Yes	✅ Yes
`Ctrl+Alt+C` Copy Masked for AI	✅ Yes	✅ Yes
`Ctrl+Alt+R` 1-Click Secret Restore	✅ Yes	✅ Yes
Auto `.gitignore` & Pre-Commit Shield	✅ Yes	✅ Yes
Shadow `.env.ai` Generation	✅ Yes	✅ Yes
Agent Audit Inspector Log	✅ Yes	✅ Yes
Client Privacy Certificate	Evaluation Watermark	Official Verified Seal + Registered License
Commercial Client & Agency Use	Personal Use	Full Commercial & Agency License

👉 Upgrade to PromptGuard Pro ($19 Lifetime / ₹499 Indian PPP)

🔒 Zero-Telemetry Architecture Guarantee

100% Client-Side: PromptGuard performs all regex scanning, Luhn validation, and token substitution locally in machine memory.
Zero Remote Analytics: No tracking pings, no telemetry beacons, no external server communication.
Offline Capable: Works completely disconnected from the internet. Your code and secrets never leave your workstation.

Developed for Developer Privacy, AI Safety, and Enterprise Security.
PromptGuard Engine v1.1.1

PromptGuard: AI Privacy & Secret Shield

Vikram Kumar

🛡️ PromptGuard: AI Prompt Privacy & Secret Shield

💎 Unlock PromptGuard Pro ($19 Lifetime / ₹499 Indian PPP)

📌 The Critical Problem: AI Prompt Credential Leaks

⚡ 100% Zero-Configuration Features (Works Out of the Box)

1. 🔍 Real-Time Inline Ghost Annotations (ErrorLens Style)

2. 🖱️ 1-Click CodeLens Buttons (Directly Above Sensitive Code)

3. 🛡️ Silent Auto-Shield Workspace Protection

4. 🔄 1-Click Secret Restore (`Ctrl+Alt+R` / `Cmd+Alt+R`)

5. 📜 Verifiable Client Privacy Audit Certificates

📋 Comprehensive 46+ Credential Detection Matrix

🚀 How to Install on Any IDE or Agent

⌨️ Everyday Keyboard Shortcuts

⭐ Free Tier vs Pro Tier

🔒 Zero-Telemetry Architecture Guarantee

PromptGuard: AI Privacy & Secret Shield

Vikram Kumar

🛡️ PromptGuard: AI Prompt Privacy & Secret Shield

💎 Unlock PromptGuard Pro ($19 Lifetime / ₹499 Indian PPP)

📌 The Critical Problem: AI Prompt Credential Leaks

⚡ 100% Zero-Configuration Features (Works Out of the Box)

1. 🔍 Real-Time Inline Ghost Annotations (ErrorLens Style)

2. 🖱️ 1-Click CodeLens Buttons (Directly Above Sensitive Code)

3. 🛡️ Silent Auto-Shield Workspace Protection

4. 🔄 1-Click Secret Restore (Ctrl+Alt+R / Cmd+Alt+R)

5. 📜 Verifiable Client Privacy Audit Certificates

📋 Comprehensive 46+ Credential Detection Matrix

🚀 How to Install on Any IDE or Agent

⌨️ Everyday Keyboard Shortcuts

⭐ Free Tier vs Pro Tier

🔒 Zero-Telemetry Architecture Guarantee

4. 🔄 1-Click Secret Restore (`Ctrl+Alt+R` / `Cmd+Alt+R`)