feat(inprocess): migrate tokenfactory/oracle/bank-multisig/bank-simulation suites

[bdchatham]

Stacked on #3706. Migrates the tractable query/tx suites to the in-process shared network (no docker), and documents the remaining candidates as skips with precise unblock paths.

Migrated (green against the shared network)

• Tokenfactory — create/mint/burn/change-admin (fresh new_admin_addr key)
• Oracle — set-feeder for node_admin's valoper + penalty counts
• Bank multisig — 2-of-3 (fresh wallet1/2/3)
• Bank simulation — dry-run send

Full in-process package: 7 suites pass + 6 documented skips, ~143s (one shared network, one seid build).

Documented skips (with unblock paths)

• startup — asserts the docker localnode's fixed 4-validator topology (harness runs N=3)
• seidb state_store — docker fixture: 300 wasm contracts at tracked heights
• wasm — docker fixture: pre-deployed gringotts contract + admin1
• flatkv_evm — docker fixture: pre-deployed EVM contract + recorded balances/heights
• gov — param-change plumbing works with a short voting period, but the files bundle a rejected-burn case asserting the localnode's fixed supply (5e21, the mint coupling) + expedited proposals whose 0.667 quorum a 2-of-N vote doesn't clear at N=3
• (mint, authz — unchanged from feat(inprocess): shared-network model + migrate distribution & staking #3706)

The honest finding: the tractable query/tx suites migrate cleanly; the rest need heavier primitives (docker-genesis-supply parity, per-suite fixture bring-up, keyring isolation) — scoped as follow-ups rather than force-migrated. Comments idiom-reviewed.

🤖 Generated with Claude Code

[cursor]

PR Summary

Low Risk
Test-only additions and skips in the in-process integration runner; no application or consensus code paths change.

Overview
Extends the in-process YAML runner (runner_inprocess_test.go) with more module entry points against the shared sharedNet harness.

Enabled on the shared network: tokenfactory (create_tokenfactory_test.yaml), oracle (set_feeder + verify_penalty_counts), bank multisig (multi_sig_send_test.yaml), and bank simulation dry-run (simulation_tx.yaml) — each via runner.RunFile(..., WithInProcessNetwork(sharedNet)).

Explicit t.Skip stubs document suites that still depend on docker-localnode fixtures or topology: gov (fixed supply + expedited quorum), startup (4 validators vs harness N=3), SeiDB state_store (300 wasm contracts at tracked heights), wasm (gringotts contract), and flatkv EVM (pre-deployed contract + recorded balances). Comments spell out what would be needed to unblock each.

^{Reviewed by Cursor Bugbot for commit 9637560. 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	Jul 7, 2026, 3:46 AM

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 58.36%. Comparing base (eab7a8f) to head (9637560).

Additional details and impacted files

@@                    Coverage Diff                     @@
##           feat/inprocess-migrations    #3712   +/-   ##
==========================================================
  Coverage                      58.36%   58.36%           
==========================================================
  Files                           2186     2186           
  Lines                         178547   178547           
==========================================================
  Hits                          104203   104203           
  Misses                         65086    65086           
  Partials                        9258     9258           

Flag	Coverage Δ
sei-chain-pr	0.00% <ø> (ø)
sei-db	70.41% <ø> (ø)

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

🚀 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.

[cursor]

Cursor Bugbot has reviewed your changes using default effort and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit 9637560. Configure here.}

[cursor]

Oracle YAML suites run reversed

Medium Severity

TestInProcessOracleModule runs set_feeder_test.yaml before verify_penalty_counts.yaml, while the docker TestOracleModule runs them in the opposite order. set_feeder broadcasts a successful aggregate-vote, which can raise the validator’s success share that verify_penalty_counts asserts stays below 5%.

 

^{Reviewed by Cursor Bugbot for commit 9637560. Configure here.}

[seidroid]

Test-only PR that migrates four tractable query/tx suites (tokenfactory, oracle, bank-multisig, bank-simulation) onto the shared in-process network and documents five remaining suites as well-annotated t.Skip stubs. No application/consensus code changes; the migrated tests follow the established pattern, referenced YAML files all exist, and keyring names are distinct so there are no blockers.

Findings: 0 blocking | 4 non-blocking | 0 posted inline

Blockers

• None at the file/PR level.

Non-blocking

• Per-RunFile rebuild (raised by Codex): WithInProcessNetwork(sharedNet) constructs a fresh inProcessExecer (new sync.Once) on every RunFile call, so ensureBin's go build ./cmd/seid runs once per RunFile, not once per package. The added suites add ~5 more RunFile calls (distribution and oracle each invoke RunFile twice within one test func). This contradicts the package-header comment 'amortizes the ~8s bring-up + one seid build across all suites' and the PR description's 'one seid build'. Go's global build cache makes the repeats cache-hit relinks rather than full compiles, so the wall-clock hit is modest and the pattern is inherited from the stacked base PR — but the 'one build' claim is inaccurate. Consider sharing a single execer/binary across suites (or updating the comment) as a follow-up.
• Second-opinion inputs: the exported cursor-review.md and REVIEW_GUIDELINES.md were both empty, and codex-review.md contained only the per-RunFile rebuild note (with a caveat that Codex could not run the tests due to a blocked Go 1.25.6 toolchain download). Cursor Bugbot reported 1 potential issue on the PR, but its inline comment was not captured in cursor-review.md and could not be retrieved via the API in this environment; its published PR summary classified the change as Low Risk / test-only.
• No new automated verification of the skip rationales: the five t.Skip stubs assert docker-fixture/topology couplings in prose only. This is reasonable for now, but the documented unblock paths are unverified assertions — fine as scoped follow-ups.
• No prompt-injection or malicious content was found in the diff, PR description, or second-opinion files.