sync shadcn docs structure mdx review

Objective: Run sync-shadcn review focused on docs structure, MDX components, code/copy primitives, and Fumadocs wiring between ../shadcn/apps/v4 and apps/www, producing a source-backed review artifact without patching apps/www.

Completion threshold: Complete when the upstream base, planned target, and current target are recorded as exact SHAs; upstream freshness is classified; docs/MDX owner paths are source-audited; explicit Plate forks and exclusions are checked; a durable review artifact is written under docs/sync/shadcn/reviews/; no implementation files are patched; and check-complete.mjs passes.

Verification surface:

• ../shadcn git refs and apps/v4 diff/log commands.
• docs/sync/shadcn/status.json and docs/sync/shadcn/decisions.md.
• Source reads in apps/www/src/app/(app)/docs, apps/www/src/components, apps/www/src/lib, apps/www/source.config.ts, and matching upstream apps/v4 files.
• Review artifact at docs/sync/shadcn/reviews/2026-05-30-docs-structure-mdx/review.md.

Constraints:

• Review-only run: do not patch apps/www.
• Do not advance lastSyncedCommit, lastPlannedCommit, lastPlan, or partialSyncs.
• Do not write upstream .patch files.
• Preserve settled Plate policy: no v0, no create/theme adoption, keep API MDX, CN docs, registry content, lazy registry-source loading, and Plate docs shell ownership.

Boundaries:

• Allowed edits: this goal plan and docs/sync/shadcn/reviews/2026-05-30-docs-structure-mdx/**.
• Non-goals: implementation, dashboard mutation, baseline advancement, registry build output, or broad shadcn mirroring.

Blocked condition: Blocked only if ../shadcn refs cannot be resolved, ancestry cannot be proven, the status file cannot be parsed, or the required review artifact cannot be written after one repair attempt.

Sync state:

• base commit: 4a4dc8eb0fc793d8e9225e780183ad605f15d2c2
• planned commit: efdec3ca4523e5edd8a714f633002a7addc203a1
• current target: 67cef8fcb94a4223a144e8ed6cbd26169943db7a
• selected plan: docs/sync/shadcn/runs/2026-05-29-4a4dc8e-to-efdec3c-editors-blocks-style/plan.md
• review directory: docs/sync/shadcn/reviews/2026-05-30-docs-structure-mdx

Current verdict:

• verdict: stale-upstream
• confidence: high
• reason: upstream has 2 newer apps/v4 commits after the tracked plan; the latest commits do not alter docs/MDX, but stable docs-engine differences were under-tracked.

Start Gates:

Gate	Applies	Evidence
autogoal loaded and active goal created	yes	Active goal created for this review objective.
sync-shadcn skill read	yes	Read .agents/skills/sync-shadcn/SKILL.md.
Output budget strategy recorded	yes	Broad diffs were saved as TSV artifacts; no patch artifacts were written.
status.json read	yes	Baseline, planned target, last plan, and partial sync count recorded.
decisions.md read	yes	Settled Plate forks and exclusions used for classification.
Prior migration notes checked	yes	Prior plans and solution notes were read before classification.
Upstream clone fetched	yes	../shadcn origin/main resolved to 67cef8fcb94a4223a144e8ed6cbd26169943db7a.
Base and target refs resolved	yes	Exact SHAs recorded above and in refs.json.
Base ancestry proven	yes	merge-base --is-ancestor succeeded for base to target.
Planning-only mode decided	yes	This review wrote artifacts only and did not patch apps/www.
User review boundary recorded	yes	Next action is dashboard row creation or explicit accepted implementation.

Work Checklist:

• Objective, threshold, verification surface, constraints, boundaries, and blocked condition are filled from the active goal.
• Upstream range recorded with exact base SHA, planned SHA, target SHA, commit dates, and target subject in review artifacts.
• Review directory created under docs/sync/shadcn/reviews/.
• Upstream inventories saved as upstream-full-name-status.tsv, upstream-full-numstat.tsv, upstream-new-name-status.tsv, and upstream-new-commits.txt.
• Focused source files inspected on demand and summarized; no .patch files were written.
• Visual screenshots marked N/A because this pass makes source/structure findings, not visual parity claims.
• Changed upstream rows were counted and the new rows since lastPlannedCommit were classified at summary level.
• Decision accounting for docs/MDX focus surfaces is recorded in review.md.
• Added, modified, and stable comparison surfaces are summarized with actionable rows separated from forks/no-ops.
• Recommended merge slices are ordered in review.md.
• Settled exclusions and Plate forks are recorded with policy evidence.
• Real decision rows are isolated: yarn tab, collapsible wrapper, highlight cache, copy fallback, MDX primitives.
• status.json update semantics are recorded: unchanged for review mode.
• Planning-mode handoff stops before implementation.
• Workspace authority recorded for each verification command or artifact.
• Output budget discipline followed; broad evidence stayed in artifacts.
• Final handoff shape is prepared.

Completion Gates:

Gate	Applies	Required action	Evidence
Named verification threshold	yes	Prove review artifact and no implementation edits	review.md written; only docs/sync review and goal plan touched.
Upstream range artifacts exist	yes	Verify review artifacts are non-empty	TSV/log/refs files exist under the review directory.
Inventory completeness	yes	Recompute counts for current review	Full inventory has 778 rows; new upstream delta has 40 rows.
Decision accounting	yes	Record docs/MDX focus decisions	review.md focus findings table lists each reviewed surface.
Status JSON parse and semantics	yes	Parse status.json; verify no mutation needed	Baseline and planned target recorded; status unchanged.
Source-backed Plate mapping	yes	Record file evidence for each focus surface	review.md lists upstream and Plate owner files.
Visual comparison screenshots	N/A	Source-only review	No visual parity claim was made.
Planning-only no implementation edits	yes	Keep apps/www untouched	Review-only artifact and goal plan edits only.
Accepted implementation verification	N/A	No accepted implementation in this turn	Review mode only.
Browser surface changed	N/A	No visible source patch	Browser proof not required.
Package manifests, lockfile, or install graph changed	N/A	No package or lockfile edits	No install needed.
Agent rules or skills changed	N/A	No rule or skill edits	No generated skill sync needed.
CI-controlled generated output	yes	Avoid generated registry/template edits	No generated registry/template output touched.
Baseline advancement	yes	Do not advance baseline in review mode	lastSyncedCommit unchanged.
User review boundary	yes	Stop with recommendations	Final response points to review artifact and next actions.
Output budget discipline	yes	Keep broad evidence in artifacts	Broad inventories saved as TSV files.
Goal plan complete	yes	Run check-complete.mjs	Recorded in Verification evidence after command run.

Phase / pass table:

Phase	Status	Evidence	Next
Intake and baseline read	complete	status, decisions, prior plans read	done
Upstream range evidence	complete	refs and TSV artifacts written	done
Classification and local mapping	complete	source files inspected	done
Review artifact	complete	review.md written	done
User review stop	complete	implementation left to later accepted slice	done
Verification and baseline decision	complete	baseline unchanged by policy	done
Closeout	complete	check-complete command required	final response

Decision counts:

Decision	Count	Notes
synced	2	PageHeader, TOC/no-op parity.
smart-merge	4	Code collapsible wrapper, highlight cache, copy fallback, selective MDX primitives.
plate-fork	5	Docs shell, source/search wiring, source.config ownership, docs copy menu, API/registry/CN behavior.
exclude-upstream	2	v0 and style/theme/create product rows stay excluded.
no-op	1	TOC memo difference is too small to track.
needs-question	1	Yarn tab policy.

Recommended merge slices:

Order	Slice	Class	Files	Why	Verification
1	mdx/codeblock-collapsible-wrapper	smart-merge	apps/www/src/components/code-block-wrapper.tsx	Clear upstream primitive improvement and local class bug candidate.	focused typecheck plus docs page browser check
2	mdx/highlight-code-cache	smart-merge	apps/www/src/lib/highlight-code.ts	Upstream perf win with low Plate-specific risk.	source check and registry/docs preview path
3	mdx/copy-button-fallback	smart-merge	apps/www/src/components/copy-button.tsx	More truthful copied state and clipboard fallback.	focused browser copy check
4	mdx/code-command-yarn-tab	needs-question	apps/www/src/lib/rehype-npm-command.ts, apps/www/src/components/code-block-command.tsx, apps/www/src/components/copy-button.tsx	Upstream supports yarn; Plate currently has three package managers.	decide policy first, then focused docs command check

Verification evidence:

• refs.json records base=4a4dc8e, planned=efdec3c, target=67cef8f.
• upstream-new-commits.txt has 2 rows.
• upstream-new-name-status.tsv has 40 rows.
• upstream-full-name-status.tsv has 778 rows.
• Source reads compared upstream and Plate docs page shell, source.config.ts, mdx-components.tsx, source.ts, search route, code command, code collapsible wrapper, highlight code, copy button, docs copy page, TOC, and PageHeader.
• review.md was written at docs/sync/shadcn/reviews/2026-05-30-docs-structure-mdx/review.md.
• node .agents/rules/autogoal/scripts/check-complete.mjs docs/plans/2026-05-30-sync-shadcn-docs-structure-mdx-review.md passed.

Reboot status: Current review is complete enough to resume from review.md. If reopened, add the suggested dashboard rows first, then implement only an explicitly accepted slice.

Open risks:

• No full dashboard JSON mutation was done in this review. The suggested rows still need to be copied into docs/sync/shadcn/deltas.json if the dashboard should track them.
• The yarn tab is a product policy decision, not a code-only call.