Slate v2 All Editor Harvest Test Processing Ralplan

status: done score: 0.96 date: 2026-05-10 skill: slate-ralplan inputs:

• docs/editor-test-harvester/lexical/report.md
• docs/editor-test-harvester/lexical/slate-processing-ledger.md
• docs/editor-test-harvester/prosemirror/report.md target repo: /Users/zbeyens/git/slate-v2

Current Verdict

Lexical is not the next broad work item. Its selected processing lane is closed: the harvest report says the selected implementation lane covered IME/history, clipboard corpus, browser transport, and table containment, and the consolidation pass found no unresolved top-level create-new, copy-now, or refactor-existing rows in that selected lane (docs/editor-test-harvester/lexical/report.md:16-20). The detailed ledger accounts for 271 inventory rows, 137 runnable portable/portable-mixed files, and 1996 indexed names, with only explicit future-owner deferrals left (docs/editor-test-harvester/lexical/slate-processing-ledger.md:13-29).

ProseMirror is the next useful pressure source. Its harvest is complete and report-only: 32 files, 848 named rows, 19 portable files, 10 portable-mixed files, 0 unresolved rows, and a clear warning not to copy the ProseMirror schema or numeric-position model (docs/editor-test-harvester/prosemirror/report.md:3-29). The high-value portable rows are DOM mutation disambiguation, composition lifecycle, clipboard context, selection geometry, decoration mapping, history rebasing, and multi-peer collaboration convergence (docs/editor-test-harvester/prosemirror/report.md:9-13).

This plan is done for the all-harvest processing milestone. PM-10 composition, PM-09 DOM-change, PM-08 collaboration convergence, PM-12 projection/widget mapping, and PM-13 geometry/RTL browser selection are processed for this milestone. PM-10 added one missing package contract for Android-style newline after composition plus one target-range test-fixture fix. PM-09 reused existing package and browser proof. PM-08 added one package convergence matrix for core replay. PM-12 added deep moved-node projection and runtime-id node-widget mapping proof. PM-13 added browser-only RTL geometry and wrapped-line selection proof. PM-01/02/03/05/06/07/11/14 were reviewed against current owner tests and explicitly deferred where the remaining rows need separate structural, clipboard, operation, history, lifecycle, browser, or device proof. No fixed or improved issue claim was added.

Intent Boundary

Intent: process every harvested editor-test source into an execution-grade Slate v2 queue without duplicating already-applied Lexical work or importing foreign editor architecture.

Outcome: a single plan that tells ralph what to execute next, what remains deferred by design, what proof command owns each slice, and when ClawSweeper has to sync claims.

In scope:

• Lexical closure accounting.
• ProseMirror PM-01 through PM-15 routing.
• Current Slate v2 owner-test grounding.
• Slate issue-ledger claim discipline.
• Future mobile/raw-device and collaboration browser deferrals.

Non-goals:

• No broad .tmp/slate-v2 implementation edits from the review pass.
• No GitHub issue comments.
• No Fixes #... claim without exact repro proof.
• No ProseMirror schema-expression, Step JSON, PluginView, NodeView, MarkView, or numeric position clone.
• No Lexical node-class, command registry, React Composer, playground product, or private MIME clone.

Decision boundaries:

• Package tests can prove model routing, history, refs, projection stores, and operation replay.
• Browser tests prove DOM selection, native composition, clipboard transport, geometry, and real rendered behavior.
• Raw mobile/device rows stay deferred until a real device lane can run bun test:mobile-device-proof:raw.
• Exact issue closure depends on current issue proof and matching browser/device coverage, not on architecture adjacency.

Decision Brief

Principles:

• Keep Slate raw and unopinionated.
• Steal behavior pressure, not foreign API shape.
• Prefer strengthening existing Slate contract files over adding duplicate test islands.
• Browser/runtime behavior must be proved where browsers own the bug.
• Keep claims narrow enough that a maintainer cannot fairly call them hype.

Main drivers:

• ProseMirror exposes runtime/browser pressure that Lexical did not fully cover.
• Lexical already closed the selected non-table lane, so rerunning it would be noisy.
• Current issue ledgers require exact distinction between Fixes, Improves, Related, and Not claimed (docs/slate-v2/ledgers/issue-coverage-matrix.md:17-31).
• The v2 sync ledger owns current manual live-issue state and has 630 live rows (docs/slate-issues/gitcrawl-v2-sync-ledger.md:8-16).

Chosen path:

• Treat Lexical as already processed except for explicit future-owner deferrals.
• Make ProseMirror PM-10, PM-09, PM-08, PM-12, and PM-13 the first execution queue.
• Batch later structural/clipboard/operation/history rows only after the browser/runtime slices stop moving the underlying proof owners.

Rejected paths:

• Copy all harvested tests: too much churn, bad architecture.
• Start with PM model/content-expression tests: lower value than runtime proof.
• Use jsdom composition as closure: known weak proof for Slate IME work.
• Claim mobile/IME closure from Chromium desktop or Playwright mobile viewport: dishonest and already forbidden by the local proof rules.

Current Live Slate Owners

The current owner scan found these relevant tests:

Owner	Current evidence	Plan use
Input strategy	.tmp/slate-v2/packages/slate-react/test/model-input-strategy-contract.test.ts:23-400	PM-09 and PM-10 package routing rows.
Editing kernel	.tmp/slate-v2/packages/slate-react/test/editing-kernel-contract.ts:39-419	Ownership/trace rows for DOM-change and composition routing.
Selection reconciler	.tmp/slate-v2/packages/slate-react/test/selection-reconciler-contract.ts:27-95	PM-10 target-range import and substitution rows.
History	.tmp/slate-v2/packages/slate-history/test/history-contract.ts:56-710	PM-07 history and composition undo rows.
Collaboration/history runtime	.tmp/slate-v2/packages/slate/test/collab-history-runtime-contract.ts:29-382	PM-08 peer convergence additions.
Projections	.tmp/slate-v2/packages/slate-react/test/projections-and-selection-contract.tsx:105-999	PM-12 moved-node projection rows.
Annotation store	.tmp/slate-v2/packages/slate-react/test/annotation-store-contract.tsx:126-824	PM-12 projection/data wake rows.
Widget layer	.tmp/slate-v2/packages/slate-react/test/widget-layer-contract.tsx:88-209	PM-12 widget boundary rows.
Browser selection	.tmp/slate-v2/packages/slate-browser/test/browser/selection.browser.test.ts:8-45	PM-13 currently thin; needs geometry/RTL browser proof.

Ecosystem Strategy Synthesis

Lexical gives Slate the issue-shaped browser regression habit: IME, paste, table, DOM repair, focus, and clipboard rows should preserve the real browser transport (docs/editor-test-harvester/lexical/report.md:108-126).

ProseMirror gives Slate the transaction/runtime authority habit: the strongest rows are not schema expressions, but ambiguous DOM-change routing, composition overlap, peer convergence, projection mapping, and selection geometry (docs/editor-test-harvester/prosemirror/report.md:52-68).

Tiptap remains a DX comparison point for extension ergonomics only. It should not drive this test-processing queue unless a future extension-command pass reopens public API shape.

React 19.2 pressure stays subscription and render-breadth oriented: projection, widget, and annotation tests must prove local subscriber wakeups instead of forcing broad editor rerenders.

Slate should deliberately diverge from both Lexical and ProseMirror on public shape. Slate owns path/range/operation/runtime behavior; Plate and apps own product policy.

All-Harvest Execution Queue

Priority	Source rows	Slate owner	Action	Fast gate	Final gate
1	PM-10 composition lifecycle plus the already-known Lexical IME proof lessons	model-input-strategy-contract.test.ts, selection-reconciler-contract.ts, dom-coverage-boundaries.test.ts, slate-browser/playwright	Add package routing for empty block, mark/cursor-wrapper, overlap cancel, rapid follow-up, and cross-paragraph composition; add one honest browser composition row before expanding.	cd .tmp/slate-v2 && bun test ./packages/slate-react/test/model-input-strategy-contract.test.ts ./packages/slate-react/test/selection-reconciler-contract.ts	cd .tmp/slate-v2 && playwright test playwright/integration/examples/dom-coverage-boundaries.test.ts --project=chromium plus bun check.
2	PM-09 DOM-change disambiguation	model-input-strategy-contract.test.ts, editing-kernel-contract.ts, generated-editing.test.ts	Add ambiguous add/remove/backspace/enter/replacement routing rows before one browser stress replay.	cd .tmp/slate-v2 && bun test ./packages/slate-react/test/model-input-strategy-contract.test.ts ./packages/slate-react/test/editing-kernel-contract.ts	Focused generated stress row plus bun check.
3	PM-08 collaboration convergence	collab-history-runtime-contract.ts	Add two/three-peer convergence for delayed local changes, simultaneous typing, deleted content, block movement, mark changes, and undo/redo.	cd .tmp/slate-v2 && bun test ./packages/slate/test/collab-history-runtime-contract.ts	Same focused test plus later slate-yjs browser proof if a browser collaboration claim is made.
4	PM-12 projection/decorator/widget mapping	projections-and-selection-contract.tsx, annotation-store-contract.tsx, widget-layer-contract.tsx	Add moved-node projection, widget side/boundary ordering, and lifecycle-cleanup rows without cloning PM NodeView/MarkView.	cd .tmp/slate-v2/packages/slate-react && bun test:vitest -- projections-and-selection-contract annotation-store-contract widget-layer-contract	cd .tmp/slate-v2 && bun check.
5	PM-13 geometry/RTL/browser selection	slate-browser/test/browser/selection.browser.test.ts, richtext.test.ts, generated-editing.test.ts	Add browser-only coords, RTL/bidi, wrapped line, block-boundary rect, atom arrow-motion rows.	cd .tmp/slate-v2 && bun --filter slate-browser test:selection	Browser proof on the specific example/stress target plus bun check.
6	PM-01/02/03/11 structural and clipboard context	clipboard-contract.ts, clipboard-boundary.ts, paste-html.test.ts, transform/operation tests	Add only missing open-edge, comment-wrapper, impossible insert/delete, and context-wrapper rows.	Focused clipboard/operation package tests.	Paste-html browser proof if external HTML transport changes.
7	PM-05/06 operation mapping and range/selection refs	operations-contract.ts, transaction-contract.ts, range-ref-contract.ts, selection-rebase-contract.ts	Add replace-around-like wrap/unwrap conflict and atom-boundary rebase rows.	Focused package tests.	bun check.
8	PM-07 history rebase	history-contract.ts, collab-history-runtime-contract.ts	Add simultaneous-edit and remote-rebase history rows only after PM-08 target shape is stable.	Focused history/collab tests.	bun check.
9	PM-14 lifecycle	transaction-contract.ts, surface-contract.tsx, editable-behavior.tsx	Strengthen only public Slate lifecycle rows; reject PM PluginView/NodeView API shape.	Focused package/vitest tests.	bun check.
10	Lexical future-owner deferrals	table model/navigation, raw mobile/device, slate-yjs browser, HR/block-void, drag/drop	Do not restart Lexical. Open only when the owner is accepted.	Owner-specific focused test.	Raw device or browser proof when claimed.

ProseMirror Row Routing

PM row	Status in this plan	Reason
PM-01 structural fit	deferred	Current transform, fragment, and operation owners already cover broad structural pressure; schema/filler parity is a future structural lane, not PM API cloning.
PM-02 open fragments	deferred	Slate clipboard is already strong; remaining context gaps belong to a dedicated clipboard import/export lane.
PM-03 DOM parse/export	deferred	Route future work through paste-html browser rows and slate-dom, not a standalone serializer clone.
PM-04 marks	mostly covered/defer	Slate should not clone mark-exclusion algebra.
PM-05 op mapping	deferred	Current operation/range-ref owners cover replay, inversion, move, split, merge, and selection rebase; wrap/unwrap conflicts are future focused work.
PM-06 selection/range refs	deferred	Existing selection/range owners are strong; atom-boundary and wrap/unwrap rows need a dedicated structural selection pass.
PM-07 history	deferred	Existing history rows are broad and PM-08 added peer convergence; remote/simultaneous history rebase is future collab/history work.
PM-08 collab convergence	priority 3	Current Slate has replay/metadata, not enough PM-grade peer convergence.
PM-09 DOM-change	priority 2	Direct browser/runtime gap.
PM-10 composition	priority 1	Highest risk and strongest local solution history.
PM-11 clipboard	deferred	Current clipboard, slate-dom, and paste-html browser rows cover the common transport families; comment-wrapper/context-wrapper rows are future clipboard proof.
PM-12 decorations/widgets	priority 4	React projection work needs moved-node/widget proof.
PM-13 geometry/RTL	priority 5	Current slate-browser selection proof is thin.
PM-14 lifecycle	deferred	Public transaction, surface, and editable lifecycle rows are covered; PM PluginView/NodeView/MarkView lifecycle is not a Slate API target.
PM-15 harness	skipped	Fixture style only.

Issue-Ledger Accounting

ClawSweeper related-issue pass for PM-10/PM-09 completed on 2026-05-10. It added a macro sync section to docs/slate-v2/ledgers/fork-issue-dossier.md and a planning sync section to docs/slate-v2/ledgers/issue-coverage-matrix.md. No fixed, improved, or PR-facing issue claim changed.

Reviewed input-runtime and DOM-change pressure:

• Existing related/non-claim/proof rows stay conservative for #6051, #6022, #5989, #5984, #5983, #5931, #5830, #5643, #5883, #4400, #5603, #5669, #4223, #3470, and #2051.
• #4466, #5493, #5974, and #5918 stay proof-needed or future-proof pressure until a matching keyboard, IME, browser, or raw-device row exists.
• #3309 routes to PM-12/PM-13 projection and browser-selection proof, not PM-10/PM-09 closure.

docs/slate-issues/gitcrawl-v2-sync-ledger.md already carries current manual classifications for these rows, so this pass did not churn individual table rows. docs/slate-v2/references/pr-description.md stays unchanged because the PR-facing claim count and proof references did not change.

Before closing any execution slice from this plan:

• Run one ClawSweeper related-issue pass for the touched surface.
• Update docs/slate-issues/gitcrawl-v2-sync-ledger.md when a slice claims, improves, reviews, or intentionally excludes a live issue or cluster.
• Update docs/slate-v2/ledgers/fork-issue-dossier.md for reviewed related issues.
• Update docs/slate-v2/ledgers/issue-coverage-matrix.md for each fixed or related issue claim.
• Update docs/slate-v2/references/pr-description.md only when the PR-facing fixed/improved/related counts or proof references change.

Likely related clusters for the first five slices:

• PM-10 / PM-09: v2-input-runtime, Android/IME/beforeinput, composition, placeholder, keyboard layout, native input event boundary.
• PM-08 / PM-07: v2-core-engine, collaboration selection/rebase, history undo selection state.
• PM-12: v2-react-runtime, decoration/projection/caret stability, async decoration, selector breadth.
• PM-13: v2-dom-selection, geometry, RTL, selection import/export, focus, atom/block boundary movement.

PM-10 Composition Execution Result

Status: complete for this all-harvest milestone.

Live source reread found existing browser coverage for the highest-risk ProseMirror composition rows:

• start, inside, end, emoji, overlap cancellation, rapid follow-up, and cross-paragraph replacement in .tmp/slate-v2/playwright/integration/examples/rendering-strategy-runtime.test.ts.
• bold/mark/cursor-wrapper and multiple formatted-node composition in .tmp/slate-v2/playwright/integration/examples/richtext.test.ts.
• decorated text composition in .tmp/slate-v2/playwright/integration/examples/highlighted-text.test.ts.
• placeholder empty-state composition in .tmp/slate-v2/playwright/integration/examples/placeholder.test.ts.
• inline-boundary composition in .tmp/slate-v2/playwright/integration/examples/mentions.test.ts.

Changes made in /Users/zbeyens/git/slate-v2:

• Added Android-style newline-after-composition package coverage in packages/slate-react/test/model-input-strategy-contract.test.ts.
• Fixed stale target-range import test fixtures in packages/slate-react/test/selection-reconciler-contract.ts by stubbing the selectable-target guard alongside toSlateRange.

Verification:

• cd /Users/zbeyens/git/slate-v2 && bun test ./packages/slate-react/test/model-input-strategy-contract.test.ts ./packages/slate-react/test/selection-reconciler-contract.ts ./packages/slate-react/test/editing-kernel-contract.ts passed: 20 tests, 31 expects.
• cd /Users/zbeyens/git/slate-v2 && bun run playwright playwright/integration/examples/rendering-strategy-runtime.test.ts --project=chromium -g "commits IME composition at the start|commits IME composition inside existing text|commits IME composition at the end|drops active IME composition|commits rapidly following IME compositions|commits cross-paragraph IME composition" passed: 8 tests.
• cd /Users/zbeyens/git/slate-v2 && bun turbo typecheck --filter=./packages/slate-react passed.
• cd /Users/zbeyens/git/slate-v2 && bunx biome check packages/slate-react/test/model-input-strategy-contract.test.ts packages/slate-react/test/selection-reconciler-contract.ts --fix passed with no fixes.

Claim policy: No Fixes #... or Improves #... promotion. Exact mobile/IME issue closure still requires matching browser or raw-device proof for the original issue.

PM-09 DOM-Change Execution Result

Status: complete for this all-harvest milestone.

Live source reread found existing coverage for the first DOM-change disambiguation set:

• target-range browser substitutions in .tmp/slate-v2/playwright/integration/examples/plaintext.test.ts.
• browser text mutations inside bold markup and ambiguous mark-boundary insertion in .tmp/slate-v2/playwright/integration/examples/richtext.test.ts.
• browser Backspace after selected text and selected ranges in .tmp/slate-v2/playwright/integration/examples/richtext.test.ts.
• package-level target-range import and command-routing rows in .tmp/slate-v2/packages/slate-react/test/selection-reconciler-contract.ts and .tmp/slate-v2/packages/slate-react/test/model-input-strategy-contract.test.ts.

Verification:

• cd /Users/zbeyens/git/slate-v2 && bun test ./packages/slate-react/test/model-input-strategy-contract.test.ts ./packages/slate-react/test/selection-reconciler-contract.ts ./packages/slate-react/test/editing-kernel-contract.ts passed.
• cd /Users/zbeyens/git/slate-v2 && bun run playwright playwright/integration/examples/plaintext.test.ts playwright/integration/examples/richtext.test.ts --project=chromium -g "applies beforeinput target ranges for browser text substitutions|syncs browser text mutations inside bold markup|resolves ambiguous browser insertion at a mark boundary|keeps caret editable after browser Backspace" passed: 5 tests.

Claim policy: No new issue promotion. This is runtime proof reuse, not exact upstream issue closure.

PM-08 Collaboration Convergence Execution Result

Status: complete for this all-harvest milestone.

Live source reread found existing collaboration coverage in .tmp/slate-v2/packages/slate/test/collab-history-runtime-contract.ts for:

• deterministic remote operation replay.
• typed remote collaboration metadata that skips local undo history.
• replace_children replay through the collaboration import path.
• remote operation replay without losing local bookmark ranges.
• local runtime-id behavior for remote remove and move operations.

The PM-08 gap was peer breadth. This pass added a focused three-peer convergence matrix in .tmp/slate-v2/packages/slate/test/collab-history-runtime-contract.ts for text insert, mark change, cross-block range delete, and block move commits. Each case verifies both remote peers converge to the source children while remote collaboration metadata keeps their local undo stacks empty.

Verification:

• cd /Users/zbeyens/git/slate-v2 && bun test ./packages/slate/test/collab-history-runtime-contract.ts passed: 8 tests.
• cd /Users/zbeyens/git/slate-v2 && bun turbo typecheck --filter=./packages/slate passed.
• cd /Users/zbeyens/git/slate-v2 && bunx biome check packages/slate/test/collab-history-runtime-contract.ts --fix fixed formatting, then the focused test and package typecheck were rerun and passed.

Claim policy: No Fixes #... or Improves #... promotion. This is package-level operation replay proof. It does not claim slate-yjs browser behavior, OT transform closure, simultaneous typing rebasing, or high-QPS remote-selection stability.

PM-12 Projection And Widget Mapping Execution Result

Status: complete for this all-harvest milestone.

Live source reread found existing projection, annotation, and widget coverage in:

• .tmp/slate-v2/packages/slate-react/test/projections-and-selection-contract.tsx for decoration sources, cross-node projection, runtime-id subscriber locality, source-bus refresh, and top-level structural path changes.
• .tmp/slate-v2/packages/slate-react/test/annotation-store-contract.tsx for annotation projection/data wakeups and changed runtime buckets.
• .tmp/slate-v2/packages/slate-react/test/widget-layer-contract.tsx for selection widget visibility and unrelated text-change isolation.
• .tmp/slate-v2/packages/slate-react/src/widget-store.ts, which exposes selection, annotation, and runtime-id node widgets. It is not a ProseMirror widget-decoration side-position API.

The PM-12 gap was structural depth and runtime-id widget mapping. This pass added:

• a nested moved-node projection row proving the same text runtime bucket follows a node moved from [0, 1, 0] to [1, 0], while unaffected runtime subscribers stay asleep.
• a runtime-id node-widget row proving a widget remains attached and silent when its node moves, then wakes exactly once when that runtime-id node is removed.

Verification:

• cd /Users/zbeyens/git/slate-v2/packages/slate-react && bun test:vitest -- projections-and-selection-contract annotation-store-contract widget-layer-contract passed: 28 tests.
• cd /Users/zbeyens/git/slate-v2 && bunx biome check packages/slate-react/test/projections-and-selection-contract.tsx packages/slate-react/test/widget-layer-contract.tsx --fix passed with no fixes.
• cd /Users/zbeyens/git/slate-v2 && bun turbo typecheck --filter=./packages/slate-react passed.

Claim policy: No Fixes #... or Improves #... promotion. PM-12 strengthens package-level projection/widget mapping proof. Exact decorated-selection browser closure, async decoration caret closure, product comment closure, and legacy decorate API parity are not claimed.

PM-13 Geometry And RTL Browser Selection Execution Result

Status: complete for this all-harvest milestone.

Live source reread found existing slate-browser selection proof in .tmp/slate-v2/packages/slate-browser/test/browser/selection.browser.test.ts for:

• simple DOM selection snapshot import/export.
• zero-width FEFF offset normalization back to editor offset zero.

The PM-13 gap was real browser geometry and RTL/wrapped-line pressure. This pass added browser-project rows for:

• RTL DOM selection offset mapping while browser geometry proves the first logical RTL character renders to the right of the last logical character.
• wrapped-line range rectangles, proving one editor selection produces multiple browser line rectangles.

Verification:

• cd /Users/zbeyens/git/slate-v2 && bun --filter slate-browser test:selection passed: 4 tests.
• cd /Users/zbeyens/git/slate-v2 && bunx biome check packages/slate-browser/test/browser/selection.browser.test.ts --fix fixed formatting, then the browser selection test was rerun and passed.
• cd /Users/zbeyens/git/slate-v2 && bun turbo typecheck --filter=./packages/slate-browser passed.

Claim policy: No Fixes #... or Improves #... promotion. PM-13 strengthens browser selection-geometry proof. Exact RTL Enter/new-line caret closure, mobile Persian scroll closure, list-direction behavior, atom arrow-motion, and decorated-text selection closure are not claimed.

PM-01/02/03/05/06/07/11/14 Gap-Fill Review Result

Status: complete for this all-harvest milestone.

The gap-fill review reread the ProseMirror report rows and current Slate v2 owner tests. It found no high-value missing row that should be added in this milestone.

PM rows	Current Slate owner evidence	Closure decision
PM-01 structural fit	.tmp/slate-v2/packages/slate/test/transforms/**, clipboard-contract.ts, and operations-contract.ts cover structural insert, delete, split, merge, wrap, unwrap, move, void, list, table, and invalid-operation pressure.	Deferred. Do not clone ProseMirror content expressions or filler policy. Future work should be a focused Slate structural-normalization lane.
PM-02/03/11 fragment, DOM parse/export, and clipboard context	clipboard-contract.ts, packages/slate-dom/test/clipboard-boundary.ts, and playwright/integration/examples/paste-html.test.ts cover fragment extraction/insertion, list/table fitting, custom fragment MIME, embedded HTML fragments, malformed payload fallback, multiline text fallback, Google Docs/Sheets, Quip, lists, tables, images, links, and generated paste/drop gauntlets.	Deferred. Remaining comment-wrapper/parser-context rows are real but belong to a dedicated clipboard/HTML import pass.
PM-05/06 operation mapping and selection/range refs	operations-contract.ts, range-ref-contract.ts, selection-rebase-contract.ts, selection-controller-contract.ts, and transform fixtures cover replay/inversion, replace_children, move_node, split/merge/remove rebasing, nested range refs, native/model selection ownership, and broad transform edge cases.	Deferred. Wrap/unwrap conflict and atom-boundary rows need a focused structural-selection pass, not all-harvest closure churn.
PM-07 history rebase	history-contract.ts covers grouping, redo clearing, selection restoration, composition undo, deleteFragment undo, blur/refocus selection, multi-block insertBreak undo, marked Enter undo, move undo, reverse joins, same-text deletes, and insertBreak undo; PM-08 added three-peer replay convergence in collab-history-runtime-contract.ts.	Deferred. Simultaneous remote history rebase needs a future collab/history lane and should not be claimed from package replay alone.
PM-14 lifecycle	transaction-contract.ts, surface-contract.tsx, and editable-behavior.tsx cover transaction publication, command middleware, extension registry slots, commit listeners, render policy, selector inventory, public host authoring, mount identity, void renderer contracts, callbacks, keydown context, and input-rule reads.	Deferred. Only public Slate lifecycle abstractions matter; PM PluginView, NodeView, and MarkView lifecycle are rejected as API targets.
PM-15 harness	ProseMirror report classifies it as harness/support only.	Skipped. No Slate implementation action.

Relevant current-source probes:

• cd /Users/zbeyens/git/slate-v2 && rg --files | rg 'paste-html|clipboard|selection-controller|transforms'
• cd /Users/zbeyens/git/slate-v2 && rg -n "(it|test)\\(" packages/slate/test/clipboard-contract.ts packages/slate-dom/test/clipboard-boundary.ts packages/slate/test/operations-contract.ts packages/slate/test/range-ref-contract.ts
• cd /Users/zbeyens/git/slate-v2 && rg -n "(it|test)\\(" packages/slate/test/selection-rebase-contract.ts packages/slate-history/test/history-contract.ts packages/slate/test/collab-history-runtime-contract.ts packages/slate/test/transaction-contract.ts
• cd /Users/zbeyens/git/slate-v2 && rg -n "(it|test)\\(" packages/slate-react/test/surface-contract.tsx packages/slate-react/test/editable-behavior.tsx playwright/integration/examples/paste-html.test.ts
• cd /Users/zbeyens/git/slate-v2 && rg -n "(it|test)\\(" packages/slate-react/test/selection-controller-contract.ts packages/slate-react/test/selection-controller-contract.test.ts

Claim policy: No PR-facing claim changed. These rows are explicitly accounted for as related test pressure, not fixed or improved issue closure.

Applicable Implementation Skill Notes

Skill lens	Status	Note
Vercel React / React runtime	applied	PM-12 must prove local projection/widget subscriber wakeups, not broad React rerenders.
performance-oracle / performance	applied	Runtime slices should avoid broad document scans and add perf rows only where the touched path is hot.
tdd	applied	Each execution slice starts with the harvested behavior row as a red/focused test where practical.
testing-review	applied	This is a value-ranked test queue, not a coverage sweep.
shadcn	skipped	No UI component design change.
react-useeffect	skipped	No effect API or hook behavior changed in this review pass.

High-Risk Pre-Mortem

Risk	Bad outcome	Guard
Composition proof lies	A jsdom or CDP row passes while real DOM composition still fails.	Use direct DOM composition mutation for mark/cursor-wrapper rows; keep raw mobile separate.
DOM-change rows duplicate repair tests	More tests without new runtime value.	Start with command-routing ambiguity, then add one browser replay only after the package row is clear.
Collaboration row overclaims slate-yjs	Core peer convergence passes, but browser collaboration remains unproved.	Keep PM-08 core-only unless a real slate-yjs/browser lane is added.
Projection row absorbs app policy	Slate starts cloning PM NodeView/MarkView or Lexical decorator APIs.	Only prove logical range projection, widget ordering, lifecycle cleanup, and React subscriber locality.
Geometry row becomes package-only	Coordinate/RTL behavior is claimed without a rendered browser.	PM-13 is browser-only except for helper normalization tests.
Issue claims drift	PR says Fixes for cluster adjacency.	ClawSweeper pass before slice closure; exact issue proof only.

Maintainer Objection Ledger

Objection	Answer
"Why ProseMirror after Lexical?"	Lexical closed the selected lane; ProseMirror exposes remaining runtime/browser pressure, especially DOM-change, composition, geometry, projection, and collab convergence.
"Are you turning Slate into ProseMirror?"	No. The plan rejects PM schema, numeric positions, Step JSON, PluginView, NodeView, and MarkView. Only behavior pressure maps into Slate path/range/operation/runtime vocabulary.
"Why not mark this done?"	This activation created the all-harvest queue. ClawSweeper sync and first PM execution slice are runnable, so the correct state is pending.
"Can we claim the mobile input issues after PM-10?"	Not unless exact Android/iOS device proof exists. Chromium desktop composition rows are useful but not raw mobile closure.
"Why not execute PM-01 first?"	Runtime/browser failures are higher release risk than structural fit rows that current Slate contracts already partially cover.

Plan Deltas From Review

• Created one all-harvest queue that separates Lexical closure from ProseMirror execution.
• Promoted PM-10, PM-09, PM-08, PM-12, and PM-13 as the first five slices.
• Demoted PM-01/02/03/05/06/07/11/14 to later gap-fill or strengthening rows.
• Kept PM-15 skipped as harness only.
• Added exact ClawSweeper/ledger obligations before any issue-facing closure.
• Completed PM-10 composition processing with focused package and browser proof.
• Fixed stale target-range selection-reconciler test fixtures exposed by the PM-10 package gate.
• Completed PM-09 DOM-change processing by verifying existing target-range, markup mutation, mark-boundary ambiguity, and Backspace browser rows.
• Completed PM-08 collaboration processing by adding three-peer core replay convergence for text, mark, range-delete, and move commits.
• Synced PM-08 collaboration issue accounting without promoting PR-facing claim counts.
• Completed PM-12 projection/widget processing by adding nested moved-node projection proof and runtime-id node-widget structural mapping proof.
• Synced PM-12 issue accounting without promoting PR-facing claim counts.
• Completed PM-13 geometry/RTL browser selection processing by adding RTL visual geometry and wrapped-line rectangle browser proof.
• Synced PM-13 issue accounting for #1498 without promoting PR-facing claim counts.
• Completed PM-01/02/03/05/06/07/11/14 gap-fill review by rereading current Slate owner tests and explicitly deferring the remaining rows to future focused lanes.
• Closed final accounting without changing PR-facing issue claims.
• Set completion done.

Pass-State Ledger

Pass	Status	Evidence added	Plan delta	Open issues	Next owner
Skill reload and boundary	complete	Reloaded slate-ralplan, editor-test-harvester, clawsweeper, ralph, goal workflow, learnings-researcher, and testing-review.	Scope locked to plan/accounting only.	none	n/a
Harvest artifact read	complete	Read Lexical report/processing ledger and ProseMirror report/index.	Split Lexical done from PM runnable queue.	none	n/a
Live Slate owner scan	complete	Grepped current Slate v2 package/browser test owners for input, history, collab, projections, widgets, and selection.	Added current owner table.	Source files can drift before execution; re-read before editing.	PM slice owner
Issue-accounting read	complete	Read live sync ledger and coverage matrix rules.	Added closure ledger obligations.	No current ledger edit because no claim changed.	ClawSweeper pass
Initial all-harvest scoring	complete	Score set to 0.86.	Pending until related issue pass and first execution slice are performed.	PM-10 not started.	ralph
ClawSweeper related issue pass	complete	Reviewed PM-10/PM-09 against gitcrawl archive, sync ledger, coverage matrix, and fork dossier; added macro sync rows.	No claim promotion; PR description unchanged.	Exact mobile/IME closure still needs matching browser or raw-device proof.	PM-10
PM-10 composition execution	complete	Added Android newline-after-composition package row, repaired target-range import fixture, and verified focused package/browser/type/lint gates in /Users/zbeyens/git/slate-v2.	PM-10 processed without issue-claim promotion.	Exact mobile/IME closure remains gated on matching original repro proof.	PM-09
PM-09 DOM-change execution	complete	Verified existing target-range substitution, markup mutation, ambiguous mark-boundary insertion, and browser Backspace rows.	PM-09 processed without issue-claim promotion.	Exact issue closure remains gated on matching original repro proof.	PM-08
PM-08 collaboration execution	complete	Added three-peer core replay convergence for text, mark, range-delete, and move commits; verified focused package test, package typecheck, and Biome formatting in /Users/zbeyens/git/slate-v2.	PM-08 processed without issue-claim promotion.	High-QPS remote selection, simultaneous typing rebasing, and browser collaboration remain future proof.	PM-12
PM-12 projection/widget execution	complete	Added nested moved-node projection proof and runtime-id node-widget structural mapping proof; verified focused package tests, package typecheck, and Biome formatting in /Users/zbeyens/git/slate-v2.	PM-12 processed without issue-claim promotion.	Exact decorated-selection browser closure, async decoration caret closure, and legacy decorate API parity remain future proof.	PM-13
PM-13 geometry/RTL browser selection execution	complete	Added browser-only RTL visual geometry and wrapped-line rectangle proof; verified focused browser selection test, package typecheck, and Biome formatting in /Users/zbeyens/git/slate-v2.	PM-13 processed without issue-claim promotion.	Exact RTL Enter/new-line caret closure, mobile RTL scroll, atom arrow motion, and decorated selection remain future proof.	gap-fill review
PM-01/02/03/05/06/07/11/14 gap-fill review	complete	Reread ProseMirror matrix rows and current Slate owner tests for structural fit, fragments, DOM parse/export, operation mapping, selection/range refs, history, clipboard, and lifecycle.	Remaining rows explicitly deferred to future focused lanes; no all-harvest implementation row added.	Future lanes remain, but this milestone has no runnable all-harvest owner left.	final ledger accounting
Final ledger accounting	complete	Plan, checkpoint, and continuation prompt synced; no PR-facing claim count changed.	Completion state can be done.	none for this milestone	n/a

Confidence Score

Dimension	Score	Evidence
Harvest accounting completeness	0.96	Lexical and ProseMirror harvests are both closed with inventories and indexes, and every PM row has a milestone decision.
Current Slate owner grounding	0.96	PM-10, PM-09, PM-08, PM-12, PM-13, and gap-fill owners were reread live before closure.
Issue-ledger readiness	0.94	PM-10/PM-09, PM-08, PM-12, and PM-13 related issue passes are recorded; gap-fill closure did not alter PR-facing claims.
Execution queue quality	0.96	PM-10, PM-08, PM-12, and PM-13 proved the queue can reuse existing owners and add only real gaps; later rows are explicitly future lanes.
Proof discipline	0.96	PM-10 has focused package/browser/type/lint proof, PM-09 has focused browser proof, PM-08 has focused package/type/lint proof, PM-12 has focused package/type/lint proof, PM-13 has focused browser/type/lint proof, and gap-fill rows are not overclaimed.
Closure readiness	0.96	No runnable all-harvest milestone row remains; future work is intentionally split into focused lanes.

Weighted score: 0.96. This is good enough to close the all-harvest milestone.

Completion Closure

Completion target is met:

• ClawSweeper ran for every touched issue-facing surface in this all-harvest milestone.
• PM-10 through PM-13 priority slices are implemented or proof-reused and verified from /Users/zbeyens/git/slate-v2.
• PM-01/02/03/05/06/07/11/14 rows have current owner tests and commands recorded, with explicit future-lane decisions.
• Issue claims in docs/slate-issues/gitcrawl-v2-sync-ledger.md, docs/slate-v2/ledgers/issue-coverage-matrix.md, docs/slate-v2/ledgers/fork-issue-dossier.md, and docs/slate-v2/references/pr-description.md remain conservative: no new Fixes or Improves claim was added.
• plate-2 completion checks verify planning artifacts only; Slate v2 runtime, package, and browser proof is recorded above with commands run from /Users/zbeyens/git/slate-v2.