feat(inprocess): shared-network model + migrate distribution & staking

[bdchatham]

In-process migration: shared-network model + first YAML-suite wave

Second wave of the in-process integration migration (builds on the merged harness, #3642).

What's here

• Keystone — node_admin operator key. Each validator's operator key is stored in its node keyring as node_admin, mirroring the docker localnode — so suites resolve the operator by name (valoper via --bech val, reward/commission queries, delegation targets) and run unchanged. provisionExtraKeys rejects an ExtraKey reusing the name (would silently overwrite the operator), with a regression test.
• Shared-network test model. The runner's in-process arm starts one inprocess.Network in TestMain and reuses it across suites — required (one network per process — process-global EVM singletons) and efficient (amortizes bring-up + the one seid build). Mirrors the docker job model: one cluster, suites sequential.
• Migrated: Bank, Distribution, Staking. Distribution's rewards and Staking's cross-node redelegate (node-0 → node-1) both exercise the operator-key seeding across distinct validators.

Measured (in-process, shared network)

Suite	Docker	In-process
Distribution	2.9 min	~30s
Staking	(in the 4.3-min Mint&Staking&Bank job)	~16s
Bank (send_funds)	—	folded into the shared run

All three + N=4 harness + guards green in one process (~87s package total incl. seid build).

Deferred (documented skips, each surfaced in go test)

• Authz — staking/generic YAMLs re-keys add grantee and rely on a passphrase-prompting keyring backend; the harness test backend takes none. Needs keyring-backend parity or per-suite key isolation.
• Mint — asserts the docker localnode's fixed genesis: both the mint-release schedule and the total supply (which folds in the full genesis allocation, not just minted tokens). Needs the localnode genesis (schedule + allocation) replicated, not the schedule alone.

Review

Independent, blinded, domain-led /xreview (per operator instruction: peer review independent of the systems engineer). Round 6 (keystone + shared-net + distribution) and Round 7 (staking + skips) — both RESOLVED. Highlights:

• sei-network-specialist (dissenter) closed a silent operator-overwrite footgun (guard + test), confirmed the cross-node staking is semantically faithful (distinct valopers, non-vacuous assertions), and caught that the mint blocker is the full genesis, not just the schedule.
• A rename-orphaned tx signer (an untouched caller) was caught by the full-package run, not the diff review.
• Ledger: bdchatham-designs test-harness arc, Rounds 6–7.

Follow-ups

• Keyring-backend parity / per-suite key isolation — unblocks the keys add-override suite class (authz).
• Localnode genesis replication — unblocks mint + likely other genesis-specific suites.
• Test residue — in-process runs write --generate-only tx files to the repo root (runner CWD = repoRoot for relative source); gitignore or write to a temp dir.
• distribution/rewards timing — reward-accrual-between-queries is tight; passed 3/4 local. Harness-side block-pacing if it flakes in CI.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 58.36%. Comparing base (8ea349f) to head (eab7a8f).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3706      +/-   ##
==========================================
- Coverage   59.31%   58.36%   -0.95%     
==========================================
  Files        2273     2186      -87     
  Lines      188364   178547    -9817     
==========================================
- Hits       111720   104203    -7517     
+ Misses      66583    65086    -1497     
+ Partials    10061     9258     -803     

Flag	Coverage Δ
sei-chain-pr	0.00% <ø> (?)
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 87 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.

[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 6, 2026, 6:31 PM