# a1 request: tx.from (EOA) per swap for the DEADKIT32 ramp pools (#650/#651)
Finding: in /opt/sw/ppdata/r6nat5_*.db, `sender` is the pool router (not a wallet): 983/985 pools show 1 sender in the first 60 s, so "who is on the other side" cannot be answered. Net ETH at the 3x cross is constant (2.30 ETH), so the ramp looks like one operator.
Ask: add column `eoa` (eth_getTransactionByHash.from) to ev for the first 120 swaps of each kit pool (~880 pools x 120 = ~105k txs; batch, free rpcpool, resumable) plus LP-remove/decrease events (#630) for the same pools.
Test on landing: (1) share of buy ETH from top-1 EOA before the 3x cross and distinct EOAs; if >=90% from one EOA and it is also the first seller/LP at the top, the ramp is an operator exit, not outside demand -> #651 must pass an outside-seller test before audit; (2) if outside EOAs >=5 per pool, #651 stands.
Free RPC returned 429 on both direct and rpcpool at 21:00 UTC, so I could not sample it myself.