rich text editor evidence benchmark

Objective: Build a comprehensive Evidence Kit benchmark lane for rich-text editors in benchmarks/editor, expanding the first Slate v2 vs legacy row into an artifact-backed benchmark matrix with explicit adapter gaps for Plate, ProseMirror, Lexical, and Tiptap.

Goal plan: docs/plans/2026-05-27-rich-text-editor-evidence-benchmark.md

Template: docs/plans/templates/major-task.md

Primary template: docs/plans/templates/major-task.md

Applied packs:

• docs (docs/plans/templates/packs/docs.md)
• browser (docs/plans/templates/packs/browser.md)

Major source:

• type: user request
• id / link: chat
• title: comprehensive rich-text editor benchmark
• decision to make: what Evidence Kit result owns rich-text editor benchmark claims, and which editor/runtime lanes are measured vs still adapter gaps
• decision criteria: current Slate v2 artifact families are imported; all local editor targets appear in coverage rows; non-measured editor comparisons are visible as gaps; Evidence Kit checks and perf docs pass

Major lane:

• lane: benchmark / performance
• output type: code, benchmark result JSON, evidence notes, generated perf docs
• implementation expected: yes
• affected packages / surfaces: benchmarks/editor/**, Evidence Kit docs/perf, this goal plan
• dominant risk: fake cross-editor numbers or duplicated Slate-only proof being mistaken for ProseMirror/Lexical/Plate/Tiptap runtime comparison

Completion threshold:

• benchmarks/results/rich-text-editors-latest.json exists and has at least 250 rows, at least 180 measured ok rows, all six local editor target roots, visible adapter-gap rows for non-Slate editors, and no missing required Slate v2 artifacts.
• The result imports every available Slate v2 artifact family named by the current Slate v2 benchmark docs or records an explicit missing optional row.
• README / iteration / source-map docs name the result file, measured families, adapter gaps, and verification command.
• Major-task closure is legal only when the decision criteria are satisfied or explicitly narrowed, facts/inference/recommendation are separated, required review or pressure passes are recorded, implementation gates are closed when code changed, and node .agents/rules/autogoal/scripts/check-complete.mjs docs/plans/2026-05-27-rich-text-editor-evidence-benchmark.md passes.

Verification surface:

• npm run bench:rich-text:check
• npm run bench:evidence
• npm run docs:perf && npm run docs:perf:check
• npm run docs:perf:search -- rich text editor slate-v2 lexical prosemirror
• node .agents/rules/autogoal/scripts/check-complete.mjs docs/plans/2026-05-27-rich-text-editor-evidence-benchmark.md

Constraints:

• Start from repo evidence before external claims.
• Keep helper stack proportional.
• Separate measured evidence, source evidence, inference, and recommendation.
• Do not execute implementation unless this major goal explicitly includes it.

Boundaries:

• Source of truth: current benchmarks/editor, .tmp/slate-v2 benchmark artifacts, and local sibling editor source roots.
• Allowed edit scope: Evidence Kit benchmark package, evidence notes, generated perf docs, and this plan.
• External sources: local clones only; no web lookup required.
• Browser surface: static Evidence Kit perf page at http://127.0.0.1:8765/index.html.
• Tracker sync: none.
• Non-goals: inventing ProseMirror/Lexical/Tiptap/Plate runtime numbers, restoring the deleted browser app benchmark lab, or running long browser integration sweeps.

Blocked condition:

• Block only if required Slate v2 artifacts are absent, Evidence Kit cannot parse the result rows, or the approved browser surface cannot be verified after generated docs are rebuilt.

Major state:

• task_type: major
• task_complexity: major
• current_phase: implementation
• current_phase_status: in_progress
• next_phase: verification
• goal_status: active

Current verdict:

• verdict: implementation in progress
• confidence: medium
• next owner: major-task
• reason: imported Slate v2 artifacts are available locally; non-Slate editor adapters remain explicit gaps

Completion rule:

• Do not call update_goal(status: complete) while any required checklist item remains unchecked. If an item does not apply, check it and add N/A: <reason>.
• Do not call update_goal(status: complete) until every completion threshold above is satisfied, final evidence is recorded, and node .agents/rules/autogoal/scripts/check-complete.mjs docs/plans/2026-05-27-rich-text-editor-evidence-benchmark.md passes.
• Do not create hook state for this goal. This file plus the active goal are the durable state.

Start Gates:

Gate	Applies	Evidence
major-task loaded	yes	/Users/zbeyens/git/plate-2/.agents/skills/major-task/SKILL.md
Active goal checked or created	yes	create_goal active objective
Source of truth read before analysis	yes	benchmarks/editor/**, docs/slate-v2 summaries, .tmp/slate-v2 artifacts
Major lane selected	yes	benchmark / performance
Decision criteria stated	yes	completion threshold above
Existing repo patterns / prior decisions checked	yes	existing Evidence Kit scripts, result rows, iteration notes, source map
Helper stack selected	yes	autogoal, major-task, Evidence Kit benchmark/perf docs guidance, docs-creator
External research decision recorded	yes	local clones only; no web research
Implementation expectation recorded	yes	benchmark importer + docs + generated perf docs
Workspace authority selected	yes	cwd /Users/zbeyens/git/plate-2/benchmarks/editor for Evidence Kit checks
Branch / PR expectation decided	yes	no PR requested
Docs pack selected	yes	plan created with docs pack
docs-creator loaded	yes	/Users/zbeyens/git/plate-2/.agents/skills/docs-creator/SKILL.md
Docs lane selected	yes	benchmark README/iteration/source-map reference docs
Target docs and nearest sibling docs read	yes	README.md, iterations/001-slate-v2-legacy-evidence.md, research/evidence-source-map.md
Docs style doctrine read	yes	docs-creator relevant sections read
Documented source owner identified	yes	benchmarks/results/rich-text-editors-latest.json
Browser pack selected	yes	plan created with browser pack
Browser route / app surface identified	yes	http://127.0.0.1:8765/index.html
Browser tool decision recorded	yes	tool_search did not expose Browser/browser-use; used curl proof against the live static route
Console/network caveat policy recorded	yes	static HTML route only; console/network inspection waived because Browser tool was unavailable

Work Checklist:

• Objective includes outcome, completion threshold, verification surface, constraints, boundaries, and blocked condition.
• Major source records source type, id/link, title, decision type, expected outcome, decision criteria, likely files/packages/surfaces, browser surface, and highest-leverage owner.
• Current state is mapped before proposing a new architecture, migration, benchmark, or plan.
• Existing repo patterns, prior decisions, and nearby implementation constraints are recorded before external research.
• External docs or source are used only where repo evidence does not settle the question, or N/A reason is recorded.
• Options, recommendation, tradeoffs, blast radius, and rejection reasons are recorded.
• Facts, inference, and recommendation are separated.
• Review or pressure lenses are selected and completed, or marked N/A with reason.
• If implementation happens, touched-surface packs cover docs, browser, package/API, or agent-native surfaces as needed.
• Workspace authority recorded: every proof command names the cwd/tool that owns the analyzed or changed behavior.
• Accepted/actionable review findings are fixed or explicitly rejected with evidence.
• Docs pack: docs lane, target docs, nearest sibling docs, and source owner are recorded.
• Docs pack: every named API, import, option, route, component, transform, demo, and preview is source-backed or marked N/A with reason.
• Docs pack: docs use current-state reference voice, not changelog voice.
• Docs pack: links, anchors, and previews target real leaf pages or are marked N/A with reason.
• Browser pack: route, interaction path, and expected visible outcome are recorded before proof.
• Browser pack: browser proof uses the repo-approved browser tool or records a blocker/waiver.
• Browser pack: console and network errors are checked or explicitly out of scope.
• Browser pack: screenshot, trace, or exact verification caveat is ready for final handoff.

Completion Gates:

Gate	Applies	Required action	Evidence
Named verification threshold	yes	Run the repo audit, benchmark, review, prototype, or artifact check named in this plan	npm run check passed in benchmarks/editor; rich result has 530 rows, 479 ok, 47 adapter-missing, no missing required artifacts
Current-state source audit	yes	Map current owner, boundaries, constraints, and affected surfaces	README.md, src/index.mjs, .tmp/slate-v2 artifact inventory, local editor source roots
Decision criteria closure	yes	Mark each criterion satisfied, narrowed, rejected, or blocked with evidence	Criteria satisfied for Slate v2 imported artifacts; explicitly narrowed for non-Slate runtime adapters as adapter gaps
Options / tradeoffs / rejection record	yes	Record viable options, chosen recommendation, and why alternatives lose	Decisions and tradeoffs section records import-and-gap matrix vs fake cross-editor numbers
Review / pressure pass	yes	Run selected reviewer/lens or record N/A with reason	Benchmark Guardian pressure applied: bad/unsupported/over-budget rows remain visible; no fake aggregate winner
Review findings closure	yes	Fix or explicitly reject accepted/actionable findings and record closure proof	First npm run check found pack budget drift; fixed by raising private lab dry-run pack budget to 800 KB
External-source audit	yes	Cite official/local clone/external sources when used, or record N/A	Local clones only: Plate, Slate v2, legacy Slate, ProseMirror, Lexical, Tiptap source roots recorded in result rows
Implementation gates	yes	If code changed, close primary-template and touched-surface gates; otherwise N/A	npm run check and pnpm exec biome check ... --fix passed
Final handoff contract	yes	Record recommendation, evidence, caveats, residual risk, and next owner	Final handoff contract section completed below
Final lint	yes	Run pnpm lint:fix or scoped equivalent when files changed	pnpm exec biome check benchmarks/editor/src/index.mjs benchmarks/editor/benchmarks/rich-text-editors-benchmark.mjs benchmarks/editor/package.json --fix passed
Goal plan complete	yes	Run node .agents/rules/autogoal/scripts/check-complete.mjs docs/plans/2026-05-27-rich-text-editor-evidence-benchmark.md	pass
Docs source-backed claim audit	yes	Verify docs claims against current source or record N/A	docs:perf:search and result JSON summaries verify result file, row counts, source roots, and adapter gaps
Docs links / routes / previews	yes	Verify leaf links, routes, anchors, and preview names or record N/A	README paths point to generated local result files; perf page served at http://127.0.0.1:8765/index.html
Docs MDX/content parser	N/A	Run pnpm --filter www build:contentlayer for MDX/content changes, or record N/A	no app MDX/content changed; Evidence Kit docs checked with npm run docs:perf:check
Plugin page specifics	N/A	For plugin pages, apply docs-creator kit/manual/API rules; otherwise N/A	no plugin page changed
Browser interaction proof	yes	Exercise the target route/interaction with the approved browser tool or record blocker	Browser tool unavailable from tool_search; curl route proof shows 589 rows and rich-text fixtures in served HTML
Browser console/network check	N/A	Record console/network state or why it is not applicable	static Evidence Kit page; Browser console unavailable this turn
Browser final proof artifact	yes	Record screenshot/trace/route proof or exact caveat	exact caveat: curl proof only, no Browser screenshot because tool was not exposed

Phase / pass table:

Phase	Status	Evidence	Next
Intake and source read	complete	read Evidence Kit package, Slate v2 docs/artifacts, local editor target roots	current-state map
Current-state map	complete	first lane had 59 docs rows; new artifact sources available under .tmp/slate-v2	options
Options and recommendation	complete	choose import-and-gap matrix over fake cross-editor numbers	review
Review / pressure pass	complete	kept adapter gaps and over-budget rows visible; rejected fake cross-editor numbers	implementation decision
Implementation or plan artifact	complete	rich-text-editors-benchmark.mjs, rich-text-editors-latest.json, docs edits, generated perf docs	verification
Verification	complete	npm run check, Biome check, docs perf check, curl route proof	closeout
Closeout	complete	final handoff contract below	final response

Findings:

• Current Evidence Kit docs had 59 rows / 20 fixtures before the broad import.
• .tmp/slate-v2 already had required artifacts for React, core, clipboard, collab, history, and issue #6038 lanes.
• Plate, legacy Slate, ProseMirror, Lexical, and Tiptap source roots exist locally.
• Non-Slate editors do not yet have equivalent runtime adapters in this lab.

Decisions and tradeoffs:

• Chosen: import all existing Slate v2 artifact families and add coverage rows for all rich-text editor targets.
• Rejected: fake ProseMirror/Lexical/Tiptap/Plate runtime numbers from source presence alone.
• Accepted caveat: optional Slate v2 transaction/current-memory artifacts can be visible as optional missing rows until those scripts emit local JSON.

Implementation notes:

• Added createRichTextEditorBenchmarkRows, Slate v2 artifact specs, workload coverage rows, and generic metric-stat normalization.
• Added benchmarks/rich-text-editors-benchmark.mjs.
• Added bench:rich-text:check and included it in bench:evidence / check.
• Added iterations/002-rich-text-editor-evidence-matrix.md.

Review fixes:

• Raised private Evidence Kit package dry-run pack budget from 350 KB to 800 KB after the comprehensive result made npm pack --dry-run 572 KB.

Error attempts:

Error / failed attempt	Count	Next different move	Resolution
npm run check failed on npm-pack over old 350 KB pack budget	1	Treat comprehensive evidence JSON as intentional package output, not noise	Raised pack budget to 800 KB and reran npm run check successfully

Verification evidence:

• npm run bench:rich-text:check in benchmarks/editor: pass, wrote 530 rows.
• npm run bench:evidence in benchmarks/editor: pass.
• npm run docs:perf && npm run docs:perf:check in benchmarks/editor: pass.
• npm run docs:perf:search -- rich text editor slate-v2 lexical prosemirror in benchmarks/editor: rich-text matrix and workload rows found.
• npm run check in benchmarks/editor: pass.
• pnpm exec biome check benchmarks/editor/src/index.mjs benchmarks/editor/benchmarks/rich-text-editors-benchmark.mjs benchmarks/editor/package.json --fix in repo root: pass, no fixes.
• curl -sSf http://127.0.0.1:8765/index.html | rg -n "Benchmark rows|Rich Text Editor|rich-text-editor": pass, served page reports 589 benchmark rows and rich-text workload fixtures.

Final handoff contract:

• Recommendation: use benchmarks/results/rich-text-editors-latest.json as the current rich-text benchmark authority.
• Confidence: high for Slate v2 artifact import coverage; low for non-Slate runtime comparison until adapters exist.
• Evidence: 530 rich-text rows, 589 total perf-doc benchmark rows, six local source roots, 47 visible adapter-gap rows.
• Tests / commands: see Verification evidence.
• Browser proof: curl proof against live static route; Browser tool unavailable.
• PR / tracker: none requested.
• Caveats: ProseMirror, Lexical, Plate, and Tiptap are source-backed targets, not measured runtime peers yet.
• Next owner: implement one non-Slate runtime adapter at a time, starting with ProseMirror or Lexical.

Timeline:

• 2026-05-27T20:52:29.725Z Major-task goal plan created.
• 2026-05-27T20:59:36Z Added rich-text editor benchmark importer, generated benchmarks/results/rich-text-editors-latest.json, and recorded first pass: 530 rows, 479 ok, 47 adapter-missing, 2 over-budget, 2 optional-missing.
• 2026-05-27T21:03:00Z Full Evidence Kit check passed; autogoal completion check passed.

Reboot status:

Question	Answer
Where am I?	Verification
Where am I going?	Verification, docs perf generation, closeout
What is the goal?	Comprehensive rich-text editor Evidence Kit benchmark matrix
What have I learned?	See Findings
What have I done?	See Timeline / Implementation notes

Open risks:

• Browser plugin may not be exposed in this turn; if so, use curl/static output proof and record the limitation.
• The current result is comprehensive for Slate v2 imported artifacts, not a finished runtime-vs-runtime comparison for ProseMirror, Lexical, Plate, or Tiptap.