[devansh-jagtap]

Summary

• Adds optional, typed classNames slots to Blog, AI Chat, CMS, and Form Builder so consumers can style built-in UI without replacing components. Classes merge with defaults, with consumer overrides winning Tailwind conflicts. Closes Add classNames field to plugin overrides for styling customization #36.
• Preserves the default Blog tag layout when no tag-list override is supplied, and keeps public-form styling through loading, error, and successful submission. Includes regression coverage, usage examples, and regenerated shadcn registry output.
• Prepares the paired @btst/stack and @btst/codegen 4.1.0 release on current main. Generated installations and playground templates select Stack 4.1.0.

Type of change

• Bug fix
• New plugin
• Feature / enhancement to an existing plugin
• Documentation
• Chore / refactor / tooling

Checklist

• pnpm build passes
• pnpm typecheck passes
• pnpm lint passes
• Tests added or updated: 1,096 Stack tests and 196 Codegen tests pass locally; all 23 styling tests pass
• Docs updated (docs/content/docs/); docs production build passes
• All three codegen-projects create successfully and pass hosted E2E tests on the merge commit, including live AI: Next.js 192 passed / 1 intentional framework skip; React Router 187 passed / 1 flaky test passed on retry / 1 intentional framework skip; TanStack 189 passed.
• New plugin: not applicable

Local Knip, 7 release-gate tests, and 8 packed-consumer harness tests also pass. The additional regression assertions reproduced seven failures before the fixes. Independent standards and specification reviews of 0a0e192cdecd932d721cba0b1017c9ad9cfe44d7 found no remaining blockers. All hosted workflows, both actual Vercel production deployments, and publication preflight passed on merge commit b22183f63c5e5729b234cbb1acff68c5b8e51531. Main E2E needed retries for existing live-AI flakiness and one runner with slow Ubuntu package downloads; no source, assertions, or release requirements were changed for those retries. Full E2E ran in GitHub Actions; local browser checks covered the generated Next.js preview at desktop and mobile sizes. No late review findings remain. Release 4.1.0 is published and its publication workflow passed. Both npm packages resolve through latest to 4.1.0 and the merge commit; fresh-cache downloads and tarball checksums are verified.

Slots

Plugin	Slots
Blog	container, postCard, postTitle, postMeta, tagsList, pagination
AI Chat	container, sidebar, messageList, message, userMessage, assistantMessage, input
CMS	container, contentTypeCard, contentTypeTitle, table, tableRow, pagination
Form Builder	container, table, tableRow, pagination, form

Form Builder's table and pagination slots cover both forms and submissions. form merges after the standalone FormRenderer.className prop and persists across its states. Blog's optional tagsList adds a flex wrapper only when supplied. Slots style the built-in components; replacement components own their styling. Form fields and submit-button primitives remain outside this change's scope.

[vercel]

@devansh-jagtap is attempting to deploy a commit to the olliethedev's Team Team on Vercel.

A member of the Team first needs to authorize it.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[olliethedev]

Reviewed the final head 0a0e192cdecd932d721cba0b1017c9ad9cfe44d7 against current main and issue #36. Fixed two findings: the default Blog tag layout must remain unchanged without an override, and FormRenderer styling must persist through loading/error/success. The new regression assertions failed before the fixes and now pass; both independent final-head reviews found no remaining blockers.

Local verification: 1,096 Stack tests, 196 Codegen tests, all 23 styling tests, typecheck, lint, build, Knip, docs production build, 7 release-gate tests, and 8 packed-consumer harness tests pass. Default/styled screenshots are attached; the generated Next.js preview also stayed within the viewport at desktop and mobile widths. Full hosted framework/consumer checks are still running.

The earlier Vercel authorization notices concern old fork commits. Both real deployments now pass on the final head via the matching upstream review branch. Copilot's quota notice did not contain a code finding; independent review is complete. No review threads remain unanswered.

[olliethedev]

Main verification update: TanStack passed all 189 tests. Next.js and React Router each hit the existing sidebar-chat smoke test: the live AI response exceeded its 30-second wait, then the built-in retry matched both the retained sidebar title and the message using the test's unscoped text selector. The trace shows the message rendered and the response still in progress; this PR does not change that test or streaming behavior. Next.js also recovered from a public-chat request timeout on its built-in retry.

Rerunning the two failed jobs once on the unchanged merge commit. Publication remains gated on successful full main checks; no test assertions or release checks have been relaxed.

[olliethedev]

Released in 4.1.0. Both @btst/stack@latest and @btst/codegen@latest now resolve to 4.1.0 from merge commit b22183f63c5e5729b234cbb1acff68c5b8e51531. The release workflow passed, and both published tarballs downloaded with a fresh npm cache and matched their integrity metadata.

All main workflows and both production deployments passed. Final E2E results: Next.js 192 passed / 1 intentional framework skip; TanStack 189 passed; React Router 187 passed / 1 flaky test passed on retry / 1 intentional framework skip. Existing live-AI flakiness required failed-job retries, and a later Next.js attempt needed an infrastructure retry for slow Ubuntu package downloads. No assertions or release gates were weakened. The PR description now contains the final verification and release links.