Refine review context on cosmos gasless tx

[masih]

Update context to avoid false positive AI review comment; see:

• Remove remaining unused legacy sei_* APIs #3927 (comment)

[cursor]

PR Summary

Low Risk
Documentation-only change to reviewer guidelines; no runtime or application code.

Overview
Extends REVIEW.md so automated reviewers stop flagging fee-less seid tx calls as broken.

The intro now mentions transaction processing alongside releases and upgrades. A new §3 explains that some Cosmos txs skip minimum-fee checks via IsTxGasless in ante (CheckAndChargeFees returns early on CheckTx/DeliverTx). It lists the gasless cases—first-time MsgAssociate (native-associate) and eligible MsgAggregateExchangeRateVote—and when missing --fees is still a real issue (non-gasless messages or gasless rules that no longer apply, e.g. re-association).

^{Reviewed by Cursor Bugbot for commit 1f39932. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

The latest Buf updates on your PR. Results from workflow Buf / buf (pull_request).

Build	Format	Lint	Breaking	Updated (UTC)
✅ passed	✅ passed	✅ passed	✅ passed	Aug 17, 2026, 11:47 AM

[seidroid]

Docs-only change adding a third "known false positive" section to the AI-reviewer guidelines about gasless Cosmos transactions. The core claims (the gasless set, the single-message MsgAssociate restriction, the 1-wei balance requirement, the non-gasless siblings) check out against app/antedecorators/gasless.go and the ante handlers, but two supporting citations are inaccurate and should be corrected before this becomes the authoritative reference reviewers are told to trust.

Findings: 0 blocking | 3 non-blocking | 2 posted inline

Blockers

• None at the file/PR level.

Non-blocking

• Codex classified this diff as a prompt-injection attempt because it instructs reviewers what not to report. I disagree: REVIEW.md is this repo's designated, maintainer-authored reviewer-guidance artifact and the review workflow loads it from the base branch by design, so updating it is normal governance rather than a directive smuggled into unrelated content. Every factual claim in the new section was verified against source before being accepted — which is the right handling either way, and is also why the two citation errors above matter: a suppression rule is only safe while the code references backing it are correct.
• 2 suggestion(s)/nit(s) flagged inline on specific lines.

[seidroid]

[suggestion] CheckAndChargeFees exists only on the CheckTx path (app/ante/cosmos_checktx.go:346, called at line 123 with isGasless). DeliverTx calls ChargeFees (app/ante/cosmos_delivertx.go:80, called at line 73), which takes no isGasless argument and performs no minGasPrices comparison at all — it unconditionally deducts whatever fee coins the tx declared. So IsTxGasless is indeed consulted by both paths (cosmos_checktx.go:96, cosmos_delivertx.go:42), but only for gas metering on the DeliverTx side; the early return described here is CheckTx-only.

Suggested rewording: IsTxGasless is consulted by both ante paths, and on CheckTx CheckAndChargeFees (app/ante/cosmos_checktx.go) returns before any fee comparison when the tx is gasless; DeliverTx never performs a minimum-fee check to begin with. Worth also noting the non-legacy chain skips the wrapped DeductFeeDecorator via GaslessDecorator (app/ante.go:80), since that — not CheckAndChargeFees — is where the skip happens there.

[seidroid]

[suggestion] associateKey (contracts/test/lib.js:306) runs seid tx evm associate-address, which does not build a MsgAssociate at all: CmdAssociateAddress (x/evm/client/cli/tx.go:78) signs an empty hash and POSTs a sei_associate JSON-RPC request, producing an EVM AssociateTx (MsgEVMTransaction with IsAssociateTx() true, handled in x/evm/ante/preprocess.go / app/ante/evm_checktx.go). Only native-associate produces the MsgAssociate that IsTxGasless classifies.

Citing it in this section conflates two independent association paths with different fee rules, and risks suppressing a genuine finding about the EVM path. Either drop the example or move it to its own note that says explicitly it is the EVM AssociateTx path and why fees don't apply there.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 58.49%. Comparing base (46671fb) to head (1f39932).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3933      +/-   ##
==========================================
- Coverage   59.51%   58.49%   -1.02%     
==========================================
  Files        2326     2230      -96     
  Lines      198890   188258   -10632     
==========================================
- Hits       118367   110130    -8237     
+ Misses      69279    67724    -1555     
+ Partials    11244    10404     -840     

Flag	Coverage Δ
sei-db	70.41% <ø> (ø)
sei-db-state-db	?

Flags with carried forward coverage won't be shown. Click here to find out more.
see 96 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.