How Cryptomining Domains Are Detected
Cryptomining abuse comes in two flavours. Browser-based cryptojacking injects JavaScript miners into web pages so that a visitor's CPU quietly mines cryptocurrency for someone else. Malware-based mining installs a persistent process on a compromised host that connects to a mining pool. Both patterns depend on reaching known infrastructure — pool endpoints, stratum servers, and script CDNs — and both are blocked by classifying that infrastructure. Our category covers more than 480,000 such domains.
Detection begins with the mining protocols themselves. Stratum, the dominant pool-communication protocol, has recognisable connection patterns, and pool domains advertise themselves through public listings, wallet-tracking services, and sandbox observation. Browser miners are identified by fingerprinting the WebAssembly and JavaScript payloads of known mining libraries such as CoinHive descendants, CryptoLoot, and their many forks, even when the code is obfuscated or self-hosted.
Because mining is fundamentally a resource-theft problem rather than a data-theft one, the signals differ from other malware. High-entropy WebAssembly modules, persistent WebSocket connections to pool proxies, and CPU-throttling logic embedded in page scripts are all indicators. We combine these behavioural fingerprints with hosting and registration reputation to catch both commercial mining services and covert cryptojacking campaigns.