Simplify retired IBC transaction rejection

[masih]

Replace keeper-backed IBC redundancy checks with a lightweight
type-URL rejection before transactions enter the mempool.

Remove obsolete IBC keeper threading from ante and CheckTx paths while
retaining MsgServer tombstones, begin-block behavior, and mounted stores.

[cursor]

PR Summary

Medium Risk
Changes mempool/CheckTx admission for any tx containing IBC messages and drops relayer redundancy filtering; behavior must stay aligned with deliver-path tombstones to avoid inconsistent rejection.

Overview
Retired IBC transactions are blocked earlier and more simply. The keeper-driven CheckTx/ante logic that simulated packet handling (RecvPacket, Ack, Timeout, UpdateClient) and rejected “all redundant packet” multisigs is removed in favor of rejecting any message whose type URL starts with /ibc. with ErrIBCDeprecated.

The IBC keeper is no longer threaded through HandlerOptions, CheckTxKeepers, CosmosCheckTxAnte, or the wasm app ante wiring; ibcante.NewAnteDecorator() is now stateless. CheckTx’s final step calls the shared ibcante.RejectMessages helper instead of the deleted CheckMessage.

Tests cover the ante decorator (IBC vs non-IBC) and assert the app message router still returns ErrIBCDeprecated for IBC msgs (e.g. MsgChannelCloseInit).

^{Reviewed by Cursor Bugbot for commit ec4d4aa. 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 25, 2026, 6:21 PM

[codecov]

Codecov Report

❌ Patch coverage is 85.71429% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 58.64%. Comparing base (d768283) to head (ec4d4aa).

Files with missing lines	Patch %	Lines
app/ante/cosmos_checktx.go	0.00%	1 Missing ⚠️
app/legacyabci/check_tx.go	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #4010      +/-   ##
==========================================
- Coverage   59.68%   58.64%   -1.04%     
==========================================
  Files        2261     2162      -99     
  Lines      194352   182680   -11672     
==========================================
- Hits       115994   107128    -8866     
+ Misses      67632    65766    -1866     
+ Partials    10726     9786     -940     

Flag	Coverage Δ
sei-chain-pr	48.93% <85.71%> (?)
sei-db	69.80% <ø> (ø)
sei-db-state-db	?

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
app/ante.go	65.51% <100.00%> (ø)
app/app.go	71.73% <ø> (-0.04%)	⬇️
sei-ibc-go/modules/core/ante/ante.go	100.00% <100.00%> (+91.11%)	⬆️
sei-wasmd/app/ante.go	61.29% <100.00%> (ø)
sei-wasmd/app/app.go	77.84% <ø> (-0.07%)	⬇️
app/ante/cosmos_checktx.go	50.30% <0.00%> (+5.40%)	⬆️
app/legacyabci/check_tx.go	0.00% <0.00%> (ø)

... and 99 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.

[seidroid]

The PR cleanly replaces the keeper-backed IBC redundancy checks with a type-URL prefix rejection and removes the now-unused IBC keeper threading; no leftover references remain and the msg-server tombstones, stores, and module wiring are preserved. One gap: the rejection only inspects top-level messages, so authz.MsgExec-wrapped IBC messages still reach the mempool.

Findings: 0 blocking | 3 non-blocking | 1 posted inline

Blockers

• None at the file/PR level.

Non-blocking

• [suggestion] AnteDecorator.AnteHandle no longer gates on ctx.IsCheckTx() || ctx.IsReCheckTx() and ignores simulate, so an IBC message that reaches DeliverTx now fails in the ante chain (last decorator, after fee deduction and sequence increment) instead of at message execution. That means no fees are charged and no sequence bump occurs for such a tx, and simulation of an IBC tx now errors. This is consistent with retiring IBC and the PR carries the app-hash-breaking label — worth an explicit confirmation that the fee/sequence accounting change is the intended consensus behavior.
• 1 suggestion(s)/nit(s) flagged inline on specific lines.
• 1 non-blocking pre-existing issue(s) listed below under pre-existing issues.

Pre-existing issues

• [suggestion] CheckAuthzContainsEvm is duplicated between app/ante/cosmos_checktx.go:569 and app/antedecorators/authz_nested_message.go:40 with identical logic and the same maxNestedMsgs cap. Two copies of the nested-authz walk means a future fix to one silently misses the other path.

[seidroid]

[suggestion] RejectMessages only walks tx.GetMsgs(), so an IBC message wrapped in authz.MsgExec has type URL /cosmos.authz.v1beta1.MsgExec and passes the prefix check. Such a tx enters the mempool and can be included in a block, only failing later at the retained MsgServer tombstone — which undercuts the stated goal of rejecting retired IBC traffic "before they enter the mempool."

The codebase already has the recursion this needs: CheckAuthzContainsEvm (app/ante/cosmos_checktx.go:569, called from the same CheckTx path at line 216, and app/antedecorators/authz_nested_message.go:40) descends through nested MsgExec with a maxNestedMsgs depth cap. Extending that walk — or mirroring it here — would close the hole at the single choke point every path passes through, and deserves a regression test with an IBC message nested inside MsgExec.