PromptShield Enterprise — AI Data Loss Prevention

Zero-leak in-browser prompt firewall and sensitive credential scanner for web-based generative AI workflows.

PromptShield is an enterprise-grade, in-browser Data Loss Prevention (DLP) and prompt firewall extension designed to protect engineering teams, developers, and corporate knowledge workers using conversational AI tools. Stop accidental leaks of API tokens, database connection strings, customer identifiers, and sensitive credentials before they leave your browser. --- ⚡ 100% In-Browser, Zero-Knowledge Architecture Unlike cloud proxy solutions that route your conversations through third-party servers, PromptShield operates entirely on your local machine using fast client-side scanning engines. • Sub-Millisecond Scanning Latency: Instant evaluation with zero noticeable typing delay. • Zero Cloud Overhead: No external language models, no remote proxies, and no prompt collection. • Air-Gapped Local Privacy: Operates in-memory with native WebCrypto AES-256-GCM encrypted local storage. --- 🔍 Comprehensive Credential & Data Detection PromptShield scans and intercepts sensitive information across multiple data categories: • Cloud & Infrastructure: Cloud access keys, cloud provider API tokens, storage connection strings, and cluster service tokens. • Developer & DevOps: Personal access tokens, version control authentication secrets, CI/CD tokens, and package registry keys. • Database Credentials: Embedded passwords and credentials inside relational, document, and key-value connection URIs. • Payment & Financial: Live payment gateway secret keys and webhook signing secrets. • Personal Identifiable Information (PII): Government identifiers, national identification numbers verified via mathematical checksums (such as Luhn and Verhoeff algorithms), and corporate email addresses. • Shannon Entropy Analysis: Statistical randomness inspection flagging unlabelled high-entropy secrets, private keys, and session tokens. --- 🛡️ Three Adaptive Protection Modes 1. Strict Block Mode: Instantly disables submission buttons and traps Enter keystrokes whenever leaked credentials or policy violations are detected. 2. Surrogate Masking Mode: Replaces detected secrets in-flight with reversible synthetic placeholder tokens so the assistant can help debug code without ever receiving real credentials. 3. Audit Mode: Unobtrusive background monitoring that automatically logs policy events to a local, encrypted audit trail for compliance verification. --- 🌐 Universal Web Chat Interface Compatibility PromptShield attaches seamlessly to standard web-based conversational AI interfaces. Its intelligent DOM observer automatically discovers active chat input textareas and editable prompt fields without requiring complex manual setup or configuration. --- 📊 Compliance Reporting & Export Generate audit-ready evidence for enterprise security standards: • SOC 2 Type II controls (CC6.1, CC6.6, CC6.7) • GDPR security requirements (Articles 32 & 33) • HIPAA technical safeguards (§ 164.312) • Export options: Structured cryptographically sealed JSON and SIEM-ready CSV logs stored solely on your device.