
Loading...

Loading...
Reference ยท Drift
Pin every economically load-bearing AIConfig parameter (stake floor, slash bps, fee split, timeouts, suspension, per-model fees) as a baseline, then see exactly what changed - so a fee or timeout shift never silently breaks your integration. Live from Mainnet.
import { LightNode, diffProtocolSnapshots } from "lightnode-sdk";
import lock from "./lightchain.lock.json"; // committed baseline
// Fail CI when a fee / timeout / slash / model changes under your app:
const changes = diffProtocolSnapshots(lock, await new LightNode("mainnet").protocolSnapshot());
if (changes.length) {
console.error("Protocol drift:", changes);
process.exit(1);
}Commit the downloaded lightchain.lock.json and diff it in CI - the build fails the day a fee, timeout, slash rate, or model changes under your app.