For 10 hours and 32 minutes, this address did almost one thing: it submitted liquidity-pool route transactions. Across all 1,037 eligible transactions in the frozen window, 1,036 asserted execute-swap-route; the remaining transaction failed before a route-level behavior could be asserted. The 852-millisecond median gap, 677-transaction longest success run, and 1,256,461-millisecond longest idle suggest bursty automated execution, but the evidence establishes neither an operator nor an identity behind the address.
A narrow repertoire
The whole-history record contains three route forms. Of the 1,036 asserted routes, 521 used roster-f0068/v1, 507 used roster-f0076/v1, and 8 used the packed e118 form. Every transaction called package revision 0x668909…f73446 alongside Sui framework revision @56, and every transaction used this address as its own gas owner; none was sponsored.
The capped latest-500 stream tells the same story. It covers transactions from 18:40:11 to 23:37:08 UTC: 499 asserted execute-swap-route, while one failed. Within that stream, 251 used roster-f0068/v1, 244 used roster-f0076/v1, and 4 used the packed form. The behavior appeared in every hour of the whole-history window, including 66 transactions in the 13:00 hour, 142 in the 14:00 hour, and 126 in the 23:00 hour. The busiest whole-history minute was 18:40 UTC, with 21 transactions.
The authored explanation describes direction-specific pool wrappers that consolidate an input coin, create an empty counterpart coin, and run an internal borrow, guard, and repayment path. That explains the repeated shape. It does not establish that an economic exchange occurred in any particular transaction.
Two transactions show the distinction clearly. t8Jhxm97 at 18:40:11 used coin::zero followed by pool_script_v2::swap_a2b; its boundary movements were +400 MAGMA, −121.397469 USDC, and −0.001553 SUI. C9gbsKTt shortly afterward merged coins, created the counterpart coin, and called swap_b2a; its boundary movements were −400 MAGMA, +121.35234 USDC, and +0.011702288 SUI. These are sender-boundary movements, not proof that either route completed an exchange or generated a return.
The single break
The only failure in the whole history—and also in the latest-500 stream—was 5bN6ZaHU at 19:54:01 UTC. It reached 0x668909…f73446::pool_script_v2::swap after preparing a MAGMA coin, then aborted. Application changes were rolled back. The recorded sender movement was only −0.0001874 SUI, with no proven MAGMA or USDC result. The address then resumed the same repertoire; the longest uninterrupted successful run in the whole history was 677 transactions.
What moved
Whole-history boundary accounting records the following, with each asset kept separate:
- SUI, gas evidence: 0.980980448 SUI in and 1.21803182 SUI out, net −0.237051372 SUI across 1,037 transactions.
- SUI, holding evidence: the same −0.237051372 SUI net. This is the same gas movement represented in another evidence category, so the two SUI figures must not be added.
- MAGMA, holding evidence: 206,000 MAGMA in and 208,400 MAGMA out, net −2,400 MAGMA across 1,036 transactions.
- USDC, holding evidence: 62,838.690672 USDC in and 61,991.115005 USDC out, net +847.575667 USDC across 1,036 transactions.
The latest-500 ledger alone records −1,200 MAGMA and +410.431041 USDC on the holding evidence, plus −0.109627568 SUI in both the holding and gas evidence. These figures describe boundary movements only. There are no prices here, and the stored route meaning explicitly does not establish an exchange or profit.
This address is best read as a highly repetitive route executor: narrow repertoire, sub-second median spacing, long successful bursts, and one isolated pool-script abort. That pattern is consistent with automated execution, while leaving the operator, intent, and economic outcome unresolved.
Further exits are t8Jhxm97, C9gbsKTt, 5bN6ZaHU, package revision 0x668909…f73446, and the two dominant route patterns roster-f0068/v1 and roster-f0076/v1.