Refine review context on cosmos gasless tx

REVIEW.md

@@ -1,10 +1,10 @@
 # Review guidelines for AI agents
 
 Repo-specific conventions for automated PR review (Codex, Cursor, Claude, and
-any other AI reviewer). The two patterns below look like bugs in isolation
-but are normal consequences of how Sei cuts releases and rolls out upgrades.
-Do not report them as findings unless a concrete additional signal (see the
-caveats under each) makes them real.
+any other AI reviewer). The patterns below look like bugs in isolation
+but are normal consequences of how Sei cuts releases, rolls out upgrades, and
+processes transactions. Do not report them as findings unless a concrete
+additional signal (see the caveats under each) makes them real.
 
 ## 1. Referencing a not-yet-created release tag is expected
 
@@ -61,3 +61,42 @@ If you believe a version gate is genuinely broken on its own logic (wrong
 comparison direction, wrong constant, wrong context field), still report
 it — this guidance only rules out the "the tag doesn't exist yet" and
 "old code might run against post-upgrade state" false positives.
+
+## 3. Some Cosmos transactions are gasless: don't require `--fees` on them
+
+Do not flag a `seid tx` invocation (in tests, scripts, or docs) as broken for
+omitting `--fees` before checking whether the message type is gasless. Sei's
+ante handlers classify certain transactions as gasless and skip minimum-fee
+validation entirely for them, so `minimum-gas-prices`-based fee arithmetic
+(`ceil(min-gas-price × gas-limit)`) does not apply.
+
+The classification lives in `IsTxGasless` (`app/antedecorators/gasless.go`)
+and is consumed by both the CheckTx and DeliverTx ante paths
+(`app/ante/cosmos_checktx.go`, `app/ante/cosmos_delivertx.go`), where
+`CheckAndChargeFees` returns before any fee comparison when the transaction
+is gasless. As of this writing the gasless set is:
+
+- a single-message `MsgAssociate` (`seid tx evm native-associate`) whose
+  sender is **not yet associated** — the common case in bootstrap helpers and
+  association tests; the sender only needs a nonzero balance (at least 1 wei),
+  not a fee, and
+- `MsgAggregateExchangeRateVote` from a validator without a vote in the
+  current window.
+
+Consequences for review:
+- A fee-less `native-associate` for a fresh (unassociated) account is
+  correct. Do not report "insufficient fees at CheckTx" for it, and do not
+  infer breakage from sibling commands that do pass `--fees` (e.g.
+  `bank send`, `associate-contract-address`) — those message types are not
+  gasless, so the asymmetry is intentional.
+- Best-effort re-association of an already-associated account is not gasless
+  and its CheckTx would reject, but helpers written as try/catch plus a
+  poll for `associated == true` (e.g. `associateKey` in
+  `contracts/test/lib.js`) don't need the transaction to land in that case.
+
+This becomes a real finding only when the diff makes a **non**-gasless
+message's invocation fee-less, or when a test strictly requires a gasless-set
+transaction to land for a sender that is already associated (or already voted)
+at the time of broadcast — then the fee path does apply and the transaction is
+rejected. When in doubt, check `IsTxGasless` for the authoritative set rather
+than reasoning from `minimum-gas-prices` alone.