# Two Mezo blocks, identical reserves: what a read-only MUSD check actually proves

> Source: <https://dev.to/livasticvm2primary/two-mezo-blocks-identical-reserves-what-a-read-only-musd-check-actually-proves-pd1>
> Published: 2026-10-23 17:04:47+00:00

What does a stablecoin dashboard prove when its numbers do not move? Not as much as a green status badge might imply.

I built [Mezo MUSD Proof Desk](https://livastic-vm2-primary-20260911.github.io/mezo-musd-proof-desk/), an independent, open-source, AI-assisted read-only checker for selected MUSD and MEZO contracts and documented pools on Mezo mainnet. It uses Python's standard library to collect block-pinned public chain evidence and publishes static JSON that anyone can inspect. It does **not** ask for a wallet connection.

Here is a small reproducible comparison—and the important limits of its conclusions.

The checker pinned contract reads to two different mainnet blocks on chain ID 31612:

| Observation | Block | Timestamp (UTC) | 
|---|---|---|
| First | `12709282` | 2026-10-23 14:49:48 | 
| Second | `12709416` | 2026-10-23 14:58:23 | 

The first block hash was `0xec725d...08528d9`, and the second was `0x5c398b...64a81d`. The full hashes, contract addresses, read method, and comparison record are in the [public source and pinned evidence](https://github.com/livastic-vm2-primary-20260911/mezo-musd-proof-desk).

For both samples, these on-chain values were unchanged:

| Item | Observed value at each block | 
|---|---|
| MUSD ERC-20 totalSupply | 29,482,016.986182972908040124 | 
| MEZO ERC-20 totalSupply | 1,138,540,989.748240837394828695 | 
| MUSD / BTC pool's MUSD reserve | 39,338.597921228749227732 MUSD | 
| MUSD / mUSDC pool's MUSD reserve | 6,852.662834954862377028 MUSD | 
| MUSD / mUSDT pool's MUSD reserve | 7,700.365872321640162678 MUSD | 

The three *selected* pools held a combined **53,891.626628505251767438 MUSD** in the sampled reserve fields. This figure is **not** total protocol TVL, dollars of executable liquidity, redeemable collateral, or a measure of MUSD backing.

Runtime-bytecode SHA-256 fingerprints for the sampled contracts and pools were also unchanged between these blocks. A byte-for-byte hash match is not a source-code audit or a security guarantee.

The proof collector follows a narrow process:

`eth_chainId` before any state read. A different chain is an error.`token0` / `token1`; read `getReserves` and MUSD `balanceOf` at that same block.`decimals` and `totalSupply` without silently assuming every token uses the same decimals.
The original repository contains **40 offline regression tests** across the collector and sample-report generator. These exercise validation and failure handling; passing them does not establish that the underlying contracts are safe.

**"The stablecoin is fully backed."** The shown ERC-20 supply and pool reserves say nothing by themselves about BTC collateral, outstanding liabilities, redemption capacity, or the peg.

**"A swap of this size would execute."** Stable-pool pricing depends on its invariant, fees, route, amount, changing balances, and other protocol details. A reserve is not a trade quote.

**"The protocol was inactive."** Two blocks with identical sampled values do not rule out events in other blocks, different contracts, bridged representations, or off-chain changes.

**"This is a security audit or yield recommendation."** No exploits, solvency model, price feed, borrow path, or vault APY were analyzed. The tool sends no transactions, pays no gas, and holds no customer assets.

That separation—*evidence observed* versus *economic inference unproven*—is the point of the exercise.

**Interested in a scoped paid QA comparison?** I can propose a human-reviewed, source-only report that compares two named public Mezo blocks for a documented MUSD contract/pool set. The **19 native Base Circle USDC** figure on the demo is my *seller quote*, not an accepted order or a claimed past sale. Scope, provenance, AI-assisted review, IP, exact chain/token/payee, legal suitability, and payment/acceptance terms must be agreed **before** any custom work. Reach me at **[onecursor1@gmail.com](mailto:onecursor1@gmail.com)**.

*Disclosure: independently operated, AI-assisted original analysis by one human owner; not affiliated with Mezo. No investor recommendation, peg/backing attestation, customer trade, commissioned audit, contest award, or verified income is implied.*