Loops HouseZygostat

NEAR HACKATHON
Links
Team
1 member- MIOwner
Millie
Overview
Problem: on October 1, 2026, NEAR Intents owed its users $3.8M that it did not hold. Price and volume charts showed nothing. The gap only exists in two on-chain numbers: what the contract credits its users, and what it actually holds.
What I built: Zygostat, a long-term investment thesis for NEAR that anyone can re-run. It tests NEAR on four kill conditions: usage (Intents volume), value capture (revenue against new token supply), solvency (credits vs holdings), and the cost of the next exploit. Each condition shows its number, its threshold, and whether it holds. Today all four hold, and if new data breaks one, the site says so.
- Run the check: your browser asks a public NEAR node for both numbers at the newest block, with no backend and no keys.
- The exploit view rebuilds October from archival state: the gap opens, peaks at $3.8M, and closes.
- History repeats the check over the past year: about 17,000 token balances, one shortfall, none since the fix.
- A risk model prices the next loss against revenue, with every assumption adjustable.
How NEAR fits: NEAR mainnet is the data source. Credits come from intents.near mt_batch_supply; holdings come from ft_balance_of and mt_batch_balance_of, read at historical blocks over JSON-RPC. Market data comes from DefiLlama and NearBlocks. A GitHub Action refreshes the data daily, and one command (node src/build.mjs --check) rebuilds and verifies every figure. The docs list each formula, source and limit. MIT licensed.