Recent

↑ ↓ to move ↵ to open esc to close

Zero-Telemetry Cryptographic & Anti-Tracking Verification Tools

Six deterministic browser-local security utilities protected by HTTP Content-Security-Policy isolation (connect-src 'none' & style-src 'self').

Workbench Isolation Architecture
The DT.gl Verification Workbench provides six client-side diagnostic sandboxes built in dependency-minimal Vanilla TypeScript and WebCrypto. Five secret-handling utilities enforce HTTP Content-Security-Policy connect-src none and style-src self headers at the Cloudflare edge, proving mathematically via browser security controls that raw calldata, addresses, entropy inputs, and local files never leave volatile client memory.

Why Policy-Based Online Decoders Are an OPSEC Liability

During incident response and pre-signing verification, engineers routinely paste raw multisig transaction hexadecimal strings, unpublished wallet addresses, BIP-39 test vectors, or internal PDFs into third-party web utilities. Standard web decoders load external analytics beacons, advertising pixels, and unpinned CDN scripts while transmitting user inputs to backend APIs, creating permanent IP-to-address correlation logs and supply-chain hijacking surfaces exemplified by the December 2023 CDN drainer injection.

Every DT.gl verification sandbox is architected so you never have to trust our server configuration alone: network isolation is enforced by your own browser kernel via immutable HTTP response headers delivered on the document load.

Complete Sandbox Security & Capability Matrix

Verification Sandbox Enforced HTTP CSP Policy Primary Cryptographic & Forensic Capabilities Paired Specification
1. Calldata & EIP-7702 Auditor connect-src 'none' 4-byte ABI selector decode, Safe operation=1 DELEGATECALL alert, local Keccak-256 safeTxHash, EIP-7702 0x04 tuple check Major Incidents Hub
2. Wallet RPC & UTXO Boltzmann Lab connect-src 'none' Multi-account JSON-RPC IP leak simulator, Bitcoin UTXO Boltzmann entropy calculator, vanity address-poisoning diff Wallet Privacy Hub
3. Seed Entropy & GF(256) Shamir Lab connect-src 'none' Dice/coin Shannon entropy & chi-square bias check, GF(256) Shamir k-of-n split/reconstruct, M-of-N vs MPC fault model On-Chain Privacy Guide
4. Data-Broker & DROP Notice Generator connect-src 'none' 14-vector RTB/CPNI/KYC physical exposure scorer, zero-upload CA DROP, CCPA/CPRA, and GDPR Art. 17 legal notice builder Data Brokers Hub
5. Browser & TLS/ECH Leak Inspector connect-src 'self' Canvas/WebGL/Audio/Sec-GPC fingerprint readout, local WebRTC ICE check, stateless /api/tls-echo, 65-param URL cleaner Encrypted DNS Hub
6. EXIF/PDF & Unicode MCP Sanitizer connect-src 'none' Browser RAM JPEG EXIF/XMP & PNG tEXt/iTXt/C2PA chunk stripper, PDF metadata inspector, Unicode U+E0000 & Bidi MCP scanner FIDO2 & MCP Guide

How to Audit Our Sandbox Headers in 30 Seconds

  1. Open any of the five offline tools above (for example, the Calldata & Signature Auditor) and open browser Developer Tools (F12 or Cmd+Option+I).
  2. Select the Network tab, refresh once, click the top HTML document request, and confirm Content-Security-Policy specifies connect-src 'none'; form-action 'none'; style-src 'self'; script-src 'self'.
  3. Clear the Network request log (or switch your workstation Wi-Fi/Ethernet completely offline) and run any preset or custom verification payload: zero outbound packets are emitted and every computation succeeds deterministically offline.

Architectural Design of Zero-Dependency Cryptographic Islands

Most browser-based developer tools import sprawling npm dependency trees (`ethers`, `web3.js`, `pdf.js`, or UI component frameworks) totaling megabytes of minified code whose transitive maintainers change over time. High dependency depth is incompatible with high-assurance cryptographic verification: a single compromised minor release in a transitive formatting utility can hook `HTMLInputElement.prototype.value` or override clipboard buffers during transaction signing.

Every interactive module inside the DT.gl Verification Workbench (`src/islands/*.ts`) adheres to four strict software supply-chain rules:

  • Zero External Runtime Dependencies: Core primitives including 24-round Keccak-f[1600] hashing, recursive-length prefix (RLP) decoding for EIP-7702 Type-0x04 envelopes, GF(256) finite-field polynomial evaluation and Lagrange interpolation for Shamir secret shares, combinatorial UTXO subset-sum partitioning for Boltzmann entropy, and binary JPEG/PNG chunk parsing are implemented directly in auditable TypeScript alongside native browser WebCrypto (crypto.subtle and crypto.getRandomValues).
  • Strict Separation of Presentation and Logic (`style-src 'self'`): Tool modules never write to element.style or inject dynamic <style> tags. Allowing 'unsafe-inline' stylesheets weakens browser defenses against CSS attribute-selector exfiltration attacks (`input[value^="0x4a"]`). Instead, every state transition toggles deterministic semantic classes compiled into external immutable stylesheets.
  • Volatile Memory Hygiene & Explicit Buffer Clearing: File sanitizers operate on ephemeral ArrayBuffer slices and revoke temporary Blob object URLs upon reset; statutory deletion notice generators build legal text strings strictly in volatile DOM state without writing to localStorage, IndexedDB, or cookies.
  • Real-World Forensic Test Vectors Built In: Rather than requiring users to hunt for sample inputs, each auditor ships with one-click historical exploit and privacy test vectors (such as the February 2025 Bybit Safe operation=1 DELEGATECALL payload, omnichain chain_id=0 EIP-7702 sweeper tuples, and Trojan Source CVE-2021-42574 Bidi/Unicode Tag prompt injections) so security teams can validate organizational runbooks immediately.

Choosing the Right Verification Utility by Operational Workflow

Before co-signing a treasury multisig transaction or approving an unfamiliar token contract, load the raw hex data or Safe transaction parameters into the Calldata & EIP-7702 Signature Auditor on a separate device from your coordinator machine to confirm operation === 0 (standard CALL) and verify the independent safeTxHash digest on your hardware signer screen. Before consolidating Bitcoin UTXOs or importing multiple accounts into a browser wallet, model link probabilities and batch JSON-RPC exposure in the Wallet RPC & UTXO Exposure Lab. When publishing screenshots, PDFs, or integrating third-party Model Context Protocol (MCP) configurations into an IDE agent, pass files and JSON descriptors through the EXIF/PDF & Invisible Unicode Sanitizer to eliminate hidden author metadata and invisible instruction tags.

Published Tool Technical Specifications (6)

Frequently Verified Technical Questions

How can I verify in my own browser that a DT.gl offline tool cannot transmit my input data?

Open your browser Developer Tools before pasting any input, navigate to the Network tab, reload the tool page, and inspect the HTTP response headers on the HTML document. You will see Content-Security-Policy containing connect-src none and form-action none, which instructs the browser engine to block every outbound network socket from that document.

Primary Reference: MDN CSP connect-src Reference

Why does the Anti-Tracking Inspector use connect-src self while the other five tools use connect-src none?

Five tools process sensitive local inputs such as calldata, wallet UTXOs, entropy rolls, legal notices, and local files, requiring zero network contact. Only the Anti-Tracking Inspector optionally queries our stateless /api/tls-echo edge endpoint on user click so you can inspect your live TLS 1.3 cipher suite and HTTP protocol negotiation.

Do DT.gl verification tools load Google Analytics or third-party CDN scripts?

No. Offline tool routes omit all analytics tags, self-host every font and script on the primary dt.gl origin, and use zero external npm runtime CDNs such as unpkg or jsDelivr, eliminating third-party supply-chain injection vectors.