fix(registry): remove install-breaking postinstall hook

[alexander-sei]

Summary

• remove the published Git submodule postinstall hook so consumers can install outside a Git checkout
• initialize registry submodules explicitly in build and coverage CI
• add a patch changeset and update repository guidance

Linear: PLT-842

Test plan

• Build @sei-js/registry
• Run all 14 registry tests
• Pack, install, and import the package with Git discovery disabled
• Validate changeset status and workspace diagnostics

Made with Cursor

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.75%. Comparing base (2896a13) to head (28aa381).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #320   +/-   ##
=======================================
  Coverage   79.75%   79.75%           
=======================================
  Files          83       83           
  Lines        1309     1309           
  Branches      217      217           
=======================================
  Hits         1044     1044           
  Misses        259      259           
  Partials        6        6           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 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]

Correct, well-scoped fix: the published postinstall hook ran git submodule update --init --recursive in every consumer's install directory, which fails outside a Git checkout (and mutates the consumer's own repo inside one), and the compiled dist already embeds the submodule JSON so consumers never needed it. The one gap is developer experience — nothing now initializes the submodules for a fresh local clone, and the root README's pnpm install → pnpm build:all instructions will fail.

Findings: 0 blocking | 6 non-blocking | 2 posted inline

Blockers

• None at the file/PR level.

Non-blocking

• Root README.md still says "To build all packages and docs, run pnpm install then pnpm build:all", which no longer works on a fresh clone now that nothing initializes the submodules. packages/registry/RUNBOOK.md documents git submodule update --init --recursive, but that's easy to miss from the repo root — worth adding the step (or git clone --recurse-submodules) to the README's Development section alongside this change.
• Consider a friendlier failure mode for the registry build: if the submodules are missing, tsc fails with Cannot find module '../../chain-registry/chain_info.json', which doesn't point at the fix. A one-line guard in the build script (or a prebuild check) that prints "run git submodule update --init --recursive" would save contributors a debugging round-trip.
• The Cursor second-opinion file (cursor-review.md) is empty — that pass produced no output, so its coverage isn't reflected here. Codex explicitly reported no material issues, matching this review.
• Verified as correct, not a finding: the changeset (@sei-js/registry patch) matches guideline 3, resolveJsonModule + outDir means tsc emits chain_info.json and the assetlist into dist, so the packed tarball is self-contained — consistent with the "pack, install, import with Git discovery disabled" item in the test plan. No prompt-injection or instruction-like content found in the diff or PR description.
• 2 suggestion(s)/nit(s) flagged inline on specific lines.

[seidroid]

[suggestion] Removing this from the published package is right — git submodule update --init --recursive fails outside a Git checkout and, when a consumer does install inside one, it operates on their repo's submodules. But it was also the only thing initializing the submodules for local development, so a fresh clone followed by pnpm install && pnpm build:all now fails on registry with Cannot find module '../../chain-registry/chain_info.json'.

The root package.json is "private": true, so its scripts are never published. Moving the hook there keeps local dev working with zero consumer impact:

"scripts": {
  "postinstall": "git submodule update --init --recursive",
  ...
}

(Guard it with git rev-parse --git-dir > /dev/null 2>&1 && if you want it to no-op in tarball/CI contexts.) Alternatively, document the step in the root README — see the separate non-blocking note.

[seidroid]

[nit] Correct and necessary now that postinstall is gone. Minor: neither entry in .gitmodules has nested submodules, so submodules: true would do and avoids the recursive walk. Both submodule URLs are public HTTPS, so this also works for fork PRs under the default GITHUB_TOKEN — no change needed there.