[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-automattic-match-page":3,"mdc-i0fz8a-key":36,"related-repo-automattic-match-page":1618,"related-org-automattic-match-page":1729},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":25,"repoUrl":26,"updatedAt":27,"license":28,"forks":29,"topics":30,"repo":31,"sourceUrl":34,"mdContent":35},"match-page","ensure visual parity for web pages","Orchestrate whole-PAGE visual parity, with a replayable parity log so the work survives a reconstruct. PHASE 0 — if \u003CoutputDir>\u002Fparity-log.json exists, REPLAY its logged intents onto the current build first (apply + verify each by looking), so prior parity work is re-applied to whatever the latest reconstruction produced. PHASE 1 — one batch pass measuring EVERY band's built design vs the SOURCE screenshot across all axes, emitting a per-band verdict. PHASE 2 — iterate match-section on the divergent bands only, worst-first, LOGGING each fix as a semantic parity-log entry, until each matches. Never hand-roll per-section work without the batch assessment; never assert a match without cropping source+built and looking. Dispatched by replicate-with-blocks\u002Fdesign-qa; calls match-section per band; writes\u002Freads parity-log.json.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"automattic","Automattic","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fautomattic.png",[12,16,19,22],{"name":13,"slug":14,"type":15},"QA","qa","tag",{"name":17,"slug":18,"type":15},"WordPress","wordpress",{"name":20,"slug":21,"type":15},"Screenshots","screenshots",{"name":23,"slug":24,"type":15},"Testing","testing",31,"https:\u002F\u002Fgithub.com\u002FAutomattic\u002Fdata-liberation-agent","2026-06-08T08:17:46.931915",null,1,[],{"repoUrl":26,"stars":25,"forks":29,"topics":32,"description":33},[],"Extract content from closed web platforms into WordPress-compatible WXR files.","https:\u002F\u002Fgithub.com\u002FAutomattic\u002Fdata-liberation-agent\u002Ftree\u002FHEAD\u002Fskills\u002Fmatch-page","---\nname: match-page\ndescription: Orchestrate whole-PAGE visual parity, with a replayable parity log so the work survives a reconstruct. PHASE 0 — if \u003CoutputDir>\u002Fparity-log.json exists, REPLAY its logged intents onto the current build first (apply + verify each by looking), so prior parity work is re-applied to whatever the latest reconstruction produced. PHASE 1 — one batch pass measuring EVERY band's built design vs the SOURCE screenshot across all axes, emitting a per-band verdict. PHASE 2 — iterate match-section on the divergent bands only, worst-first, LOGGING each fix as a semantic parity-log entry, until each matches. Never hand-roll per-section work without the batch assessment; never assert a match without cropping source+built and looking. Dispatched by replicate-with-blocks\u002Fdesign-qa; calls match-section per band; writes\u002Freads parity-log.json.\ndisable-model-invocation: true\nallowed-tools:\n  - Bash\n  - Read\n  - Edit\n  - Agent\n  - mcp__plugin_playwright_playwright__browser_navigate\n  - mcp__plugin_playwright_playwright__browser_resize\n  - mcp__plugin_playwright_playwright__browser_evaluate\n---\n\n## Why this exists\n\nReaching design parity by improvising per-section edits is slow, inconsistent, and prone to runaways. The fix is a fixed two-phase shape: **measure the whole page once, THEN iterate only the sections that diverge.** Phase 1 is cheap (one render + one DOM measurement); it tells you exactly which sections need work and on which axes, so Phase 2 is targeted, not exploratory.\n\n## Inputs\n\n`outputDir` (e.g. `~\u002FStudio\u002F_liberations\u002Fexample.com`), `studioSitePath`, `themeSlug`, page `slug` + `sourceUrl`, `previewUrl` (e.g. `http:\u002F\u002Flocalhost:8881`; front page renders at `\u002F`). Captured specs: `outputDir\u002Fsections\u002F\u003Cslug>.json` → `sections[]` (ordered by `top`; each carries `top`, `height`, `backgroundColor`, `layout.padTopPx\u002FpadBottomPx\u002Fgap`, `fullBleed`, `headingSizes`\u002F`headingLineHeights`\u002F`headingFamilies`, `bodyTextSizes`\u002F`bodyLineHeights`\u002F`bodyFamilies`, `cells`, `textAlign`, plus `styledHtml` — the computed-style oracle). Source screenshot: `outputDir\u002Fscreenshots\u002Fdesktop\u002F\u003Cslug>.png` (read its real width with `identify`; compare at THAT width).\n- `replicaShotsDir` — the replica screenshots directory holding `comparison.json` (v2) and the `diff\u002F` artifacts; used by the Phase-1 parity gate.\n\n**CRITICAL — compare against the SOURCE SCREENSHOT, not the captured numbers.** The captured `SectionSpec` values (`headingSizes`, `padTopPx`, `height`, …) are measured at the **1440px desktop capture**; the page renders and is compared at the source screenshot width (often **1008px**), and the deterministic emitter scales type\u002Fspacing by `vw`\u002F`clamp`. So built `26px` is a captured `36px` scaled at 1008 — NOT a divergence. Diffing built-render against captured-@1440 numbers produces all false positives and causes thrash. The TRUTH for visual parity is **built render vs source render at the same width** (pixels vs pixels). Use the captured numbers only as the values to APPLY in Phase 2, never as the comparison target.\n\n## The parity log — `\u003CoutputDir>\u002Fparity-log.json`\n\nA replayable record of the SEMANTIC steps to reach parity. NOT literal markup diffs (those break when a reconstruction changes structure) — each entry is an *intent* an AI can re-apply by looking at the source. Format + helpers: `src\u002Flib\u002Freplicate\u002Fparity-log.ts` (`parseParityLog`, `serializeParityLog`, `upsertEntry`, `replayableEntries`). One entry per band×axes:\n\n```json\n{ \"version\": 1, \"page\": \"homepage\", \"sourceUrl\": \"https:\u002F\u002F…\u002F\",\n  \"entries\": [\n    { \"band\": \"services\", \"bandIndex\": 1,\n      \"divergence\": \"card titles render stacked above the cards; cards show only the number\",\n      \"sourceTarget\": \"each card = number + title (large) + description + Read More\",\n      \"fix\": \"place each section heading into its card as the title at the captured size; drop the orphaned stacked headings\",\n      \"axes\": [\"structure\", \"type-size\"], \"status\": \"applied\",\n      \"promoteToEmitter\": \"extractor: capture the card title into the cell, not the number; merge a split 2-line title\",\n      \"evidence\": { \"source\": \"screenshots\u002Fparity\u002Fservices-src.png\", \"built\": \"screenshots\u002Fparity\u002Fservices-built.png\" } }\n  ] }\n```\n\n`status`: `applied` (fix in place + verified), `residual` (couldn't fully close — note why), `pending` (identified, not yet applied). `promoteToEmitter` is the backlog: a recurring, rule-shaped divergence belongs in the extractor\u002Femitter — once encoded there, DELETE the entry from the log (it no longer needs replay). Write the log with `upsertEntry` semantics: re-polishing the same band×axes UPDATES its entry; a new band\u002Faxis APPENDS.\n\n## PHASE 0 — Replay the parity log (if it exists)\n\nIf `\u003CoutputDir>\u002Fparity-log.json` is present, run BEFORE the fresh assessment: for each `replayableEntries` entry (applied\u002Fresidual, in band order), apply its `fix` to the CURRENT build (the band may have just been regenerated by a reconstruct), then VERIFY by cropping that band from the source screenshot + the built page and reading BOTH — never assume the replay worked. Update the entry's `status` from what you see. Replay re-grounds on the source each time, so it survives structural drift. Then continue to Phase 1, which catches anything the log didn't cover.\n\n## PHASE 1 — Batch assessment (ALWAYS do this first; one pass)\n\n**Parity gate (run FIRST, before any per-section work):** read `\u003CreplicaShotsDir>\u002Fcomparison.json` (version 2). For each page: if every scored viewport has `status:\"ok\"`, `score >= 0.995` (PARITY_GATE_SCORE), AND `heightMismatchRatio \u003C= 0.02` (HEIGHT_MISMATCH_THRESHOLD), the page is ALREADY AT PARITY — skip it entirely (no section assessment, no match-section dispatch) and log `gate-skip: \u003Cslug> score=\u003Cs> hmr=\u003Cr>`. A high score with `heightMismatchRatio > 0.02` does NOT pass — that combination is the short-capture artifact; treat the height mismatch itself as a high-severity finding and proceed with assessment. When the file is version 1 (no mismatch fields), the gate is score-only — never assume missing fields mean zero mismatch. If comparison.json is absent (compare never ran), there is no gate — assess every page. When `fullPageScore` is present AND the heights match (`originHeight === replicaHeight`), prefer it over the crop score for the gate decision (same 0.995 bar) — it sees below the fold. When the heights differ, gate on the crop score + `heightMismatchRatio` as above: `fullPageScore` counts the magenta padding as diff by construction, so even a benign sub-threshold height delta (the 32px admin-bar artifact ≈ 0.6% on a long page) caps it below 0.995 and would make the gate-skip permanently unreachable.\n\n1. Read the source screenshot width: `identify -format \"%w\" outputDir\u002Fscreenshots\u002Fdesktop\u002F\u003Cslug>.png`. Render the built page ONCE at THAT width: `npx tsx scripts\u002F_shot.ts \"\u003CpreviewUrl>\u002F\u003Cslug>\u002F?v=1\" outputDir\u002Fscreenshots\u002Fbuilt-\u003Cslug>.png \u003CsrcWidth>` (it scrolls to settle lazy images; do not use the Playwright MCP screenshot — it times out on tall pages).\n2. Get the built band boundaries: `browser_navigate` at `\u003CsrcWidth>×900`, then `browser_evaluate` to list the band container's children (the `.entry-content`\u002F`.wp-block-post-content` children — a MIX of `\u003Cdiv class=\"wp-block-cover\">` heroes and `\u003Csection class=\"wp-block-group\">` bands) with each one's `getBoundingClientRect()` top + height. That gives N built bands in document order.\n3. For EACH band, crop the SAME vertical range from BOTH the source screenshot and the built screenshot (`magick \u003Cimg> -crop \u003CW>x\u003CH>+0+\u003Ctop> +repage out.png`; source band tops scale from the built tops by the page-height ratio, or detect bg-color band boundaries) and READ the two crops side by side. Flag the band DIVERGENT on the axes you can SEE: background color, heading size\u002Fweight relative to the band, body text size, alignment (centered vs left), inter-element spacing, full-bleed vs boxed, dropped media\u002Fcontent, button style. This is a visual judgement per band, not a number diff.\n4. **Print the per-band verdict table** (band → divergent axes, or DONE). Bands that look identical are DONE — do not touch them.\n\nThis whole phase is ONE render + N band crops. It replaces eyeballing the whole page and guessing, and it does NOT chase captured @1440 numbers.\n\n**Diff-image hygiene:** a `.diff.png` \u002F `.padded.png` marks WHERE pixels differ — never read a color, font, or size from it. Red is the difference marker; magenta (in `.padded.png`) means one side has no content at that location (layout overrun — high severity). Read every actual value from the SOURCE screenshot.\n\n## PHASE 2 — Iterate divergent sections (worst-first)\n\nFor each divergent section, ordered by composite descending:\n\n1. Run **`match-section`** for that section index (dispatch a subagent pointed at `skills\u002Fmatch-section\u002FSKILL.md`, OR apply inline if you are already that executor): apply the captured values for the divergent axes (from the spec \u002F `styledHtml`) as canonicalization-safe core block attributes, render, and eyes-on compare source-crop vs built-crop until visually identical.\n2. **Re-measure ONLY that section** (re-run the Phase-1 measurement for its index) to confirm its divergent cells cleared. Do not re-render\u002Fre-measure the whole page each time.\n3. **Log the fix.** Once the band verifies as matched (or residual), upsert a parity-log entry for it — `{band, bandIndex, divergence, sourceTarget, fix, axes, status, evidence}` — and, if the divergence is a recurring rule-shaped gap (e.g. a capture\u002Frender defect any site would hit), set `promoteToEmitter` with how to encode it. Save `\u003CoutputDir>\u002Fparity-log.json` after each band so a crash\u002Finterrupt never loses the record. Also save the source + built evidence crops under `\u003CoutputDir>\u002Fscreenshots\u002Fparity\u002F`.\n4. **Bounds (anti-runaway):** at most 3 fix cycles per band; if a band still diverges after 3, log it `residual` (with the remaining axes + why) and move on — do NOT keep grinding one band. Cap the whole page at a sane total (e.g. ≤ 2× the band count of cycles); if hit, stop and report.\n\nMechanics for applying + rendering are in `skills\u002Fmatch-section\u002FSKILL.md` (BACK UP post_content to a file first; post_content via `studio wp post update`; EXACT unique-string edits, never a greedy regex; mirror into both theme pattern copies; `cache flush`). Preserve ALL content; never touch the nav template, footer part, or a band already marked DONE.\n\n## Output\n\n- The Phase-1 verdict table (before): every band, divergent axes flagged.\n- Per divergent band: the axes fixed (before→after) and final MATCHED\u002FRESIDUAL.\n- The verdict table re-run at the end (after), confirmed by looking at the crops.\n- Total cycles spent (so a runaway is visible).\n- **`\u003CoutputDir>\u002Fparity-log.json`** written with one entry per fix (the replayable record).\n- **Promote-to-emitter backlog:** the list of `promoteToEmitter` notes — the recurring, rule-shaped divergences that should graduate into the extractor\u002Femitter (and then leave the log).\n- Before declaring done any page where at least one section was dispatched to match-section: `liberate_refine_report {outputDir, slug}` must pass for that page. A failed coverage check reopens the page — dispatch match-section to fix the accounting. Pages with zero match-section dispatches (gate-skipped, or all bands matched at assessment) have no refine\u002F\u003Cslug>\u002F directory and are exempt — do NOT call the tool for them.\n\nNever report the page MATCHED without having cropped source+built per band and looked at both. The parity log is what makes this sustainable: a future reconstruct (done for its own reason) regenerates the deterministic baseline, Phase 0 replays the log onto it, and anything that has since been encoded into the emitter is simply dropped from the log.\n",{"data":37,"body":47},{"name":4,"description":6,"disable-model-invocation":38,"allowed-tools":39},true,[40,41,42,43,44,45,46],"Bash","Read","Edit","Agent","mcp__plugin_playwright_playwright__browser_navigate","mcp__plugin_playwright_playwright__browser_resize","mcp__plugin_playwright_playwright__browser_evaluate",{"type":48,"children":49},"root",[50,59,73,79,292,324,414,426,476,1056,1103,1109,1142,1148,1236,1352,1357,1390,1396,1401,1502,1530,1536,1607,1612],{"type":51,"tag":52,"props":53,"children":55},"element","h2",{"id":54},"why-this-exists",[56],{"type":57,"value":58},"text","Why this exists",{"type":51,"tag":60,"props":61,"children":62},"p",{},[63,65,71],{"type":57,"value":64},"Reaching design parity by improvising per-section edits is slow, inconsistent, and prone to runaways. The fix is a fixed two-phase shape: ",{"type":51,"tag":66,"props":67,"children":68},"strong",{},[69],{"type":57,"value":70},"measure the whole page once, THEN iterate only the sections that diverge.",{"type":57,"value":72}," Phase 1 is cheap (one render + one DOM measurement); it tells you exactly which sections need work and on which axes, so Phase 2 is targeted, not exploratory.",{"type":51,"tag":52,"props":74,"children":76},{"id":75},"inputs",[77],{"type":57,"value":78},"Inputs",{"type":51,"tag":60,"props":80,"children":81},{},[82,89,91,97,99,105,107,113,115,121,123,129,130,136,137,143,145,151,153,159,161,167,169,175,177,182,183,189,190,196,197,203,204,210,211,217,218,224,225,231,232,238,239,245,246,252,253,259,260,266,268,274,276,282,284,290],{"type":51,"tag":83,"props":84,"children":86},"code",{"className":85},[],[87],{"type":57,"value":88},"outputDir",{"type":57,"value":90}," (e.g. ",{"type":51,"tag":83,"props":92,"children":94},{"className":93},[],[95],{"type":57,"value":96},"~\u002FStudio\u002F_liberations\u002Fexample.com",{"type":57,"value":98},"), ",{"type":51,"tag":83,"props":100,"children":102},{"className":101},[],[103],{"type":57,"value":104},"studioSitePath",{"type":57,"value":106},", ",{"type":51,"tag":83,"props":108,"children":110},{"className":109},[],[111],{"type":57,"value":112},"themeSlug",{"type":57,"value":114},", page ",{"type":51,"tag":83,"props":116,"children":118},{"className":117},[],[119],{"type":57,"value":120},"slug",{"type":57,"value":122}," + ",{"type":51,"tag":83,"props":124,"children":126},{"className":125},[],[127],{"type":57,"value":128},"sourceUrl",{"type":57,"value":106},{"type":51,"tag":83,"props":131,"children":133},{"className":132},[],[134],{"type":57,"value":135},"previewUrl",{"type":57,"value":90},{"type":51,"tag":83,"props":138,"children":140},{"className":139},[],[141],{"type":57,"value":142},"http:\u002F\u002Flocalhost:8881",{"type":57,"value":144},"; front page renders at ",{"type":51,"tag":83,"props":146,"children":148},{"className":147},[],[149],{"type":57,"value":150},"\u002F",{"type":57,"value":152},"). Captured specs: ",{"type":51,"tag":83,"props":154,"children":156},{"className":155},[],[157],{"type":57,"value":158},"outputDir\u002Fsections\u002F\u003Cslug>.json",{"type":57,"value":160}," → ",{"type":51,"tag":83,"props":162,"children":164},{"className":163},[],[165],{"type":57,"value":166},"sections[]",{"type":57,"value":168}," (ordered by ",{"type":51,"tag":83,"props":170,"children":172},{"className":171},[],[173],{"type":57,"value":174},"top",{"type":57,"value":176},"; each carries ",{"type":51,"tag":83,"props":178,"children":180},{"className":179},[],[181],{"type":57,"value":174},{"type":57,"value":106},{"type":51,"tag":83,"props":184,"children":186},{"className":185},[],[187],{"type":57,"value":188},"height",{"type":57,"value":106},{"type":51,"tag":83,"props":191,"children":193},{"className":192},[],[194],{"type":57,"value":195},"backgroundColor",{"type":57,"value":106},{"type":51,"tag":83,"props":198,"children":200},{"className":199},[],[201],{"type":57,"value":202},"layout.padTopPx\u002FpadBottomPx\u002Fgap",{"type":57,"value":106},{"type":51,"tag":83,"props":205,"children":207},{"className":206},[],[208],{"type":57,"value":209},"fullBleed",{"type":57,"value":106},{"type":51,"tag":83,"props":212,"children":214},{"className":213},[],[215],{"type":57,"value":216},"headingSizes",{"type":57,"value":150},{"type":51,"tag":83,"props":219,"children":221},{"className":220},[],[222],{"type":57,"value":223},"headingLineHeights",{"type":57,"value":150},{"type":51,"tag":83,"props":226,"children":228},{"className":227},[],[229],{"type":57,"value":230},"headingFamilies",{"type":57,"value":106},{"type":51,"tag":83,"props":233,"children":235},{"className":234},[],[236],{"type":57,"value":237},"bodyTextSizes",{"type":57,"value":150},{"type":51,"tag":83,"props":240,"children":242},{"className":241},[],[243],{"type":57,"value":244},"bodyLineHeights",{"type":57,"value":150},{"type":51,"tag":83,"props":247,"children":249},{"className":248},[],[250],{"type":57,"value":251},"bodyFamilies",{"type":57,"value":106},{"type":51,"tag":83,"props":254,"children":256},{"className":255},[],[257],{"type":57,"value":258},"cells",{"type":57,"value":106},{"type":51,"tag":83,"props":261,"children":263},{"className":262},[],[264],{"type":57,"value":265},"textAlign",{"type":57,"value":267},", plus ",{"type":51,"tag":83,"props":269,"children":271},{"className":270},[],[272],{"type":57,"value":273},"styledHtml",{"type":57,"value":275}," — the computed-style oracle). Source screenshot: ",{"type":51,"tag":83,"props":277,"children":279},{"className":278},[],[280],{"type":57,"value":281},"outputDir\u002Fscreenshots\u002Fdesktop\u002F\u003Cslug>.png",{"type":57,"value":283}," (read its real width with ",{"type":51,"tag":83,"props":285,"children":287},{"className":286},[],[288],{"type":57,"value":289},"identify",{"type":57,"value":291},"; compare at THAT width).",{"type":51,"tag":293,"props":294,"children":295},"ul",{},[296],{"type":51,"tag":297,"props":298,"children":299},"li",{},[300,306,308,314,316,322],{"type":51,"tag":83,"props":301,"children":303},{"className":302},[],[304],{"type":57,"value":305},"replicaShotsDir",{"type":57,"value":307}," — the replica screenshots directory holding ",{"type":51,"tag":83,"props":309,"children":311},{"className":310},[],[312],{"type":57,"value":313},"comparison.json",{"type":57,"value":315}," (v2) and the ",{"type":51,"tag":83,"props":317,"children":319},{"className":318},[],[320],{"type":57,"value":321},"diff\u002F",{"type":57,"value":323}," artifacts; used by the Phase-1 parity gate.",{"type":51,"tag":60,"props":325,"children":326},{},[327,332,334,340,342,347,348,354,355,360,362,367,369,374,376,382,383,389,391,397,399,405,407,412],{"type":51,"tag":66,"props":328,"children":329},{},[330],{"type":57,"value":331},"CRITICAL — compare against the SOURCE SCREENSHOT, not the captured numbers.",{"type":57,"value":333}," The captured ",{"type":51,"tag":83,"props":335,"children":337},{"className":336},[],[338],{"type":57,"value":339},"SectionSpec",{"type":57,"value":341}," values (",{"type":51,"tag":83,"props":343,"children":345},{"className":344},[],[346],{"type":57,"value":216},{"type":57,"value":106},{"type":51,"tag":83,"props":349,"children":351},{"className":350},[],[352],{"type":57,"value":353},"padTopPx",{"type":57,"value":106},{"type":51,"tag":83,"props":356,"children":358},{"className":357},[],[359],{"type":57,"value":188},{"type":57,"value":361},", …) are measured at the ",{"type":51,"tag":66,"props":363,"children":364},{},[365],{"type":57,"value":366},"1440px desktop capture",{"type":57,"value":368},"; the page renders and is compared at the source screenshot width (often ",{"type":51,"tag":66,"props":370,"children":371},{},[372],{"type":57,"value":373},"1008px",{"type":57,"value":375},"), and the deterministic emitter scales type\u002Fspacing by ",{"type":51,"tag":83,"props":377,"children":379},{"className":378},[],[380],{"type":57,"value":381},"vw",{"type":57,"value":150},{"type":51,"tag":83,"props":384,"children":386},{"className":385},[],[387],{"type":57,"value":388},"clamp",{"type":57,"value":390},". So built ",{"type":51,"tag":83,"props":392,"children":394},{"className":393},[],[395],{"type":57,"value":396},"26px",{"type":57,"value":398}," is a captured ",{"type":51,"tag":83,"props":400,"children":402},{"className":401},[],[403],{"type":57,"value":404},"36px",{"type":57,"value":406}," scaled at 1008 — NOT a divergence. Diffing built-render against captured-@1440 numbers produces all false positives and causes thrash. The TRUTH for visual parity is ",{"type":51,"tag":66,"props":408,"children":409},{},[410],{"type":57,"value":411},"built render vs source render at the same width",{"type":57,"value":413}," (pixels vs pixels). Use the captured numbers only as the values to APPLY in Phase 2, never as the comparison target.",{"type":51,"tag":52,"props":415,"children":417},{"id":416},"the-parity-log-outputdirparity-logjson",[418,420],{"type":57,"value":419},"The parity log — ",{"type":51,"tag":83,"props":421,"children":423},{"className":422},[],[424],{"type":57,"value":425},"\u003CoutputDir>\u002Fparity-log.json",{"type":51,"tag":60,"props":427,"children":428},{},[429,431,437,439,445,447,453,454,460,461,467,468,474],{"type":57,"value":430},"A replayable record of the SEMANTIC steps to reach parity. NOT literal markup diffs (those break when a reconstruction changes structure) — each entry is an ",{"type":51,"tag":432,"props":433,"children":434},"em",{},[435],{"type":57,"value":436},"intent",{"type":57,"value":438}," an AI can re-apply by looking at the source. Format + helpers: ",{"type":51,"tag":83,"props":440,"children":442},{"className":441},[],[443],{"type":57,"value":444},"src\u002Flib\u002Freplicate\u002Fparity-log.ts",{"type":57,"value":446}," (",{"type":51,"tag":83,"props":448,"children":450},{"className":449},[],[451],{"type":57,"value":452},"parseParityLog",{"type":57,"value":106},{"type":51,"tag":83,"props":455,"children":457},{"className":456},[],[458],{"type":57,"value":459},"serializeParityLog",{"type":57,"value":106},{"type":51,"tag":83,"props":462,"children":464},{"className":463},[],[465],{"type":57,"value":466},"upsertEntry",{"type":57,"value":106},{"type":51,"tag":83,"props":469,"children":471},{"className":470},[],[472],{"type":57,"value":473},"replayableEntries",{"type":57,"value":475},"). One entry per band×axes:",{"type":51,"tag":477,"props":478,"children":483},"pre",{"className":479,"code":480,"language":481,"meta":482,"style":482},"language-json shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","{ \"version\": 1, \"page\": \"homepage\", \"sourceUrl\": \"https:\u002F\u002F…\u002F\",\n  \"entries\": [\n    { \"band\": \"services\", \"bandIndex\": 1,\n      \"divergence\": \"card titles render stacked above the cards; cards show only the number\",\n      \"sourceTarget\": \"each card = number + title (large) + description + Read More\",\n      \"fix\": \"place each section heading into its card as the title at the captured size; drop the orphaned stacked headings\",\n      \"axes\": [\"structure\", \"type-size\"], \"status\": \"applied\",\n      \"promoteToEmitter\": \"extractor: capture the card title into the cell, not the number; merge a split 2-line title\",\n      \"evidence\": { \"source\": \"screenshots\u002Fparity\u002Fservices-src.png\", \"built\": \"screenshots\u002Fparity\u002Fservices-built.png\" } }\n  ] }\n","json","",[484],{"type":51,"tag":83,"props":485,"children":486},{"__ignoreMap":482},[487,599,626,695,734,772,810,905,943,1043],{"type":51,"tag":488,"props":489,"children":491},"span",{"class":490,"line":29},"line",[492,498,503,509,514,519,525,530,534,539,543,547,551,557,561,565,569,573,577,581,585,590,594],{"type":51,"tag":488,"props":493,"children":495},{"style":494},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[496],{"type":57,"value":497},"{",{"type":51,"tag":488,"props":499,"children":500},{"style":494},[501],{"type":57,"value":502}," \"",{"type":51,"tag":488,"props":504,"children":506},{"style":505},"--shiki-light:#9C3EDA;--shiki-default:#C792EA;--shiki-dark:#C792EA",[507],{"type":57,"value":508},"version",{"type":51,"tag":488,"props":510,"children":511},{"style":494},[512],{"type":57,"value":513},"\"",{"type":51,"tag":488,"props":515,"children":516},{"style":494},[517],{"type":57,"value":518},":",{"type":51,"tag":488,"props":520,"children":522},{"style":521},"--shiki-light:#F76D47;--shiki-default:#F78C6C;--shiki-dark:#F78C6C",[523],{"type":57,"value":524}," 1",{"type":51,"tag":488,"props":526,"children":527},{"style":494},[528],{"type":57,"value":529},",",{"type":51,"tag":488,"props":531,"children":532},{"style":494},[533],{"type":57,"value":502},{"type":51,"tag":488,"props":535,"children":536},{"style":505},[537],{"type":57,"value":538},"page",{"type":51,"tag":488,"props":540,"children":541},{"style":494},[542],{"type":57,"value":513},{"type":51,"tag":488,"props":544,"children":545},{"style":494},[546],{"type":57,"value":518},{"type":51,"tag":488,"props":548,"children":549},{"style":494},[550],{"type":57,"value":502},{"type":51,"tag":488,"props":552,"children":554},{"style":553},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[555],{"type":57,"value":556},"homepage",{"type":51,"tag":488,"props":558,"children":559},{"style":494},[560],{"type":57,"value":513},{"type":51,"tag":488,"props":562,"children":563},{"style":494},[564],{"type":57,"value":529},{"type":51,"tag":488,"props":566,"children":567},{"style":494},[568],{"type":57,"value":502},{"type":51,"tag":488,"props":570,"children":571},{"style":505},[572],{"type":57,"value":128},{"type":51,"tag":488,"props":574,"children":575},{"style":494},[576],{"type":57,"value":513},{"type":51,"tag":488,"props":578,"children":579},{"style":494},[580],{"type":57,"value":518},{"type":51,"tag":488,"props":582,"children":583},{"style":494},[584],{"type":57,"value":502},{"type":51,"tag":488,"props":586,"children":587},{"style":553},[588],{"type":57,"value":589},"https:\u002F\u002F…\u002F",{"type":51,"tag":488,"props":591,"children":592},{"style":494},[593],{"type":57,"value":513},{"type":51,"tag":488,"props":595,"children":596},{"style":494},[597],{"type":57,"value":598},",\n",{"type":51,"tag":488,"props":600,"children":602},{"class":490,"line":601},2,[603,608,613,617,621],{"type":51,"tag":488,"props":604,"children":605},{"style":494},[606],{"type":57,"value":607},"  \"",{"type":51,"tag":488,"props":609,"children":610},{"style":505},[611],{"type":57,"value":612},"entries",{"type":51,"tag":488,"props":614,"children":615},{"style":494},[616],{"type":57,"value":513},{"type":51,"tag":488,"props":618,"children":619},{"style":494},[620],{"type":57,"value":518},{"type":51,"tag":488,"props":622,"children":623},{"style":494},[624],{"type":57,"value":625}," [\n",{"type":51,"tag":488,"props":627,"children":629},{"class":490,"line":628},3,[630,635,639,645,649,653,657,662,666,670,674,679,683,687,691],{"type":51,"tag":488,"props":631,"children":632},{"style":494},[633],{"type":57,"value":634},"    {",{"type":51,"tag":488,"props":636,"children":637},{"style":494},[638],{"type":57,"value":502},{"type":51,"tag":488,"props":640,"children":642},{"style":641},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[643],{"type":57,"value":644},"band",{"type":51,"tag":488,"props":646,"children":647},{"style":494},[648],{"type":57,"value":513},{"type":51,"tag":488,"props":650,"children":651},{"style":494},[652],{"type":57,"value":518},{"type":51,"tag":488,"props":654,"children":655},{"style":494},[656],{"type":57,"value":502},{"type":51,"tag":488,"props":658,"children":659},{"style":553},[660],{"type":57,"value":661},"services",{"type":51,"tag":488,"props":663,"children":664},{"style":494},[665],{"type":57,"value":513},{"type":51,"tag":488,"props":667,"children":668},{"style":494},[669],{"type":57,"value":529},{"type":51,"tag":488,"props":671,"children":672},{"style":494},[673],{"type":57,"value":502},{"type":51,"tag":488,"props":675,"children":676},{"style":641},[677],{"type":57,"value":678},"bandIndex",{"type":51,"tag":488,"props":680,"children":681},{"style":494},[682],{"type":57,"value":513},{"type":51,"tag":488,"props":684,"children":685},{"style":494},[686],{"type":57,"value":518},{"type":51,"tag":488,"props":688,"children":689},{"style":521},[690],{"type":57,"value":524},{"type":51,"tag":488,"props":692,"children":693},{"style":494},[694],{"type":57,"value":598},{"type":51,"tag":488,"props":696,"children":698},{"class":490,"line":697},4,[699,704,709,713,717,721,726,730],{"type":51,"tag":488,"props":700,"children":701},{"style":494},[702],{"type":57,"value":703},"      \"",{"type":51,"tag":488,"props":705,"children":706},{"style":641},[707],{"type":57,"value":708},"divergence",{"type":51,"tag":488,"props":710,"children":711},{"style":494},[712],{"type":57,"value":513},{"type":51,"tag":488,"props":714,"children":715},{"style":494},[716],{"type":57,"value":518},{"type":51,"tag":488,"props":718,"children":719},{"style":494},[720],{"type":57,"value":502},{"type":51,"tag":488,"props":722,"children":723},{"style":553},[724],{"type":57,"value":725},"card titles render stacked above the cards; cards show only the number",{"type":51,"tag":488,"props":727,"children":728},{"style":494},[729],{"type":57,"value":513},{"type":51,"tag":488,"props":731,"children":732},{"style":494},[733],{"type":57,"value":598},{"type":51,"tag":488,"props":735,"children":737},{"class":490,"line":736},5,[738,742,747,751,755,759,764,768],{"type":51,"tag":488,"props":739,"children":740},{"style":494},[741],{"type":57,"value":703},{"type":51,"tag":488,"props":743,"children":744},{"style":641},[745],{"type":57,"value":746},"sourceTarget",{"type":51,"tag":488,"props":748,"children":749},{"style":494},[750],{"type":57,"value":513},{"type":51,"tag":488,"props":752,"children":753},{"style":494},[754],{"type":57,"value":518},{"type":51,"tag":488,"props":756,"children":757},{"style":494},[758],{"type":57,"value":502},{"type":51,"tag":488,"props":760,"children":761},{"style":553},[762],{"type":57,"value":763},"each card = number + title (large) + description + Read More",{"type":51,"tag":488,"props":765,"children":766},{"style":494},[767],{"type":57,"value":513},{"type":51,"tag":488,"props":769,"children":770},{"style":494},[771],{"type":57,"value":598},{"type":51,"tag":488,"props":773,"children":775},{"class":490,"line":774},6,[776,780,785,789,793,797,802,806],{"type":51,"tag":488,"props":777,"children":778},{"style":494},[779],{"type":57,"value":703},{"type":51,"tag":488,"props":781,"children":782},{"style":641},[783],{"type":57,"value":784},"fix",{"type":51,"tag":488,"props":786,"children":787},{"style":494},[788],{"type":57,"value":513},{"type":51,"tag":488,"props":790,"children":791},{"style":494},[792],{"type":57,"value":518},{"type":51,"tag":488,"props":794,"children":795},{"style":494},[796],{"type":57,"value":502},{"type":51,"tag":488,"props":798,"children":799},{"style":553},[800],{"type":57,"value":801},"place each section heading into its card as the title at the captured size; drop the orphaned stacked headings",{"type":51,"tag":488,"props":803,"children":804},{"style":494},[805],{"type":57,"value":513},{"type":51,"tag":488,"props":807,"children":808},{"style":494},[809],{"type":57,"value":598},{"type":51,"tag":488,"props":811,"children":813},{"class":490,"line":812},7,[814,818,823,827,831,836,840,845,849,853,857,862,866,871,875,880,884,888,892,897,901],{"type":51,"tag":488,"props":815,"children":816},{"style":494},[817],{"type":57,"value":703},{"type":51,"tag":488,"props":819,"children":820},{"style":641},[821],{"type":57,"value":822},"axes",{"type":51,"tag":488,"props":824,"children":825},{"style":494},[826],{"type":57,"value":513},{"type":51,"tag":488,"props":828,"children":829},{"style":494},[830],{"type":57,"value":518},{"type":51,"tag":488,"props":832,"children":833},{"style":494},[834],{"type":57,"value":835}," [",{"type":51,"tag":488,"props":837,"children":838},{"style":494},[839],{"type":57,"value":513},{"type":51,"tag":488,"props":841,"children":842},{"style":553},[843],{"type":57,"value":844},"structure",{"type":51,"tag":488,"props":846,"children":847},{"style":494},[848],{"type":57,"value":513},{"type":51,"tag":488,"props":850,"children":851},{"style":494},[852],{"type":57,"value":529},{"type":51,"tag":488,"props":854,"children":855},{"style":494},[856],{"type":57,"value":502},{"type":51,"tag":488,"props":858,"children":859},{"style":553},[860],{"type":57,"value":861},"type-size",{"type":51,"tag":488,"props":863,"children":864},{"style":494},[865],{"type":57,"value":513},{"type":51,"tag":488,"props":867,"children":868},{"style":494},[869],{"type":57,"value":870},"],",{"type":51,"tag":488,"props":872,"children":873},{"style":494},[874],{"type":57,"value":502},{"type":51,"tag":488,"props":876,"children":877},{"style":641},[878],{"type":57,"value":879},"status",{"type":51,"tag":488,"props":881,"children":882},{"style":494},[883],{"type":57,"value":513},{"type":51,"tag":488,"props":885,"children":886},{"style":494},[887],{"type":57,"value":518},{"type":51,"tag":488,"props":889,"children":890},{"style":494},[891],{"type":57,"value":502},{"type":51,"tag":488,"props":893,"children":894},{"style":553},[895],{"type":57,"value":896},"applied",{"type":51,"tag":488,"props":898,"children":899},{"style":494},[900],{"type":57,"value":513},{"type":51,"tag":488,"props":902,"children":903},{"style":494},[904],{"type":57,"value":598},{"type":51,"tag":488,"props":906,"children":908},{"class":490,"line":907},8,[909,913,918,922,926,930,935,939],{"type":51,"tag":488,"props":910,"children":911},{"style":494},[912],{"type":57,"value":703},{"type":51,"tag":488,"props":914,"children":915},{"style":641},[916],{"type":57,"value":917},"promoteToEmitter",{"type":51,"tag":488,"props":919,"children":920},{"style":494},[921],{"type":57,"value":513},{"type":51,"tag":488,"props":923,"children":924},{"style":494},[925],{"type":57,"value":518},{"type":51,"tag":488,"props":927,"children":928},{"style":494},[929],{"type":57,"value":502},{"type":51,"tag":488,"props":931,"children":932},{"style":553},[933],{"type":57,"value":934},"extractor: capture the card title into the cell, not the number; merge a split 2-line title",{"type":51,"tag":488,"props":936,"children":937},{"style":494},[938],{"type":57,"value":513},{"type":51,"tag":488,"props":940,"children":941},{"style":494},[942],{"type":57,"value":598},{"type":51,"tag":488,"props":944,"children":946},{"class":490,"line":945},9,[947,951,956,960,964,969,973,978,982,986,990,995,999,1003,1007,1012,1016,1020,1024,1029,1033,1038],{"type":51,"tag":488,"props":948,"children":949},{"style":494},[950],{"type":57,"value":703},{"type":51,"tag":488,"props":952,"children":953},{"style":641},[954],{"type":57,"value":955},"evidence",{"type":51,"tag":488,"props":957,"children":958},{"style":494},[959],{"type":57,"value":513},{"type":51,"tag":488,"props":961,"children":962},{"style":494},[963],{"type":57,"value":518},{"type":51,"tag":488,"props":965,"children":966},{"style":494},[967],{"type":57,"value":968}," {",{"type":51,"tag":488,"props":970,"children":971},{"style":494},[972],{"type":57,"value":502},{"type":51,"tag":488,"props":974,"children":975},{"style":521},[976],{"type":57,"value":977},"source",{"type":51,"tag":488,"props":979,"children":980},{"style":494},[981],{"type":57,"value":513},{"type":51,"tag":488,"props":983,"children":984},{"style":494},[985],{"type":57,"value":518},{"type":51,"tag":488,"props":987,"children":988},{"style":494},[989],{"type":57,"value":502},{"type":51,"tag":488,"props":991,"children":992},{"style":553},[993],{"type":57,"value":994},"screenshots\u002Fparity\u002Fservices-src.png",{"type":51,"tag":488,"props":996,"children":997},{"style":494},[998],{"type":57,"value":513},{"type":51,"tag":488,"props":1000,"children":1001},{"style":494},[1002],{"type":57,"value":529},{"type":51,"tag":488,"props":1004,"children":1005},{"style":494},[1006],{"type":57,"value":502},{"type":51,"tag":488,"props":1008,"children":1009},{"style":521},[1010],{"type":57,"value":1011},"built",{"type":51,"tag":488,"props":1013,"children":1014},{"style":494},[1015],{"type":57,"value":513},{"type":51,"tag":488,"props":1017,"children":1018},{"style":494},[1019],{"type":57,"value":518},{"type":51,"tag":488,"props":1021,"children":1022},{"style":494},[1023],{"type":57,"value":502},{"type":51,"tag":488,"props":1025,"children":1026},{"style":553},[1027],{"type":57,"value":1028},"screenshots\u002Fparity\u002Fservices-built.png",{"type":51,"tag":488,"props":1030,"children":1031},{"style":494},[1032],{"type":57,"value":513},{"type":51,"tag":488,"props":1034,"children":1035},{"style":494},[1036],{"type":57,"value":1037}," }",{"type":51,"tag":488,"props":1039,"children":1040},{"style":494},[1041],{"type":57,"value":1042}," }\n",{"type":51,"tag":488,"props":1044,"children":1046},{"class":490,"line":1045},10,[1047,1052],{"type":51,"tag":488,"props":1048,"children":1049},{"style":494},[1050],{"type":57,"value":1051},"  ]",{"type":51,"tag":488,"props":1053,"children":1054},{"style":494},[1055],{"type":57,"value":1042},{"type":51,"tag":60,"props":1057,"children":1058},{},[1059,1064,1066,1071,1073,1079,1081,1087,1089,1094,1096,1101],{"type":51,"tag":83,"props":1060,"children":1062},{"className":1061},[],[1063],{"type":57,"value":879},{"type":57,"value":1065},": ",{"type":51,"tag":83,"props":1067,"children":1069},{"className":1068},[],[1070],{"type":57,"value":896},{"type":57,"value":1072}," (fix in place + verified), ",{"type":51,"tag":83,"props":1074,"children":1076},{"className":1075},[],[1077],{"type":57,"value":1078},"residual",{"type":57,"value":1080}," (couldn't fully close — note why), ",{"type":51,"tag":83,"props":1082,"children":1084},{"className":1083},[],[1085],{"type":57,"value":1086},"pending",{"type":57,"value":1088}," (identified, not yet applied). ",{"type":51,"tag":83,"props":1090,"children":1092},{"className":1091},[],[1093],{"type":57,"value":917},{"type":57,"value":1095}," is the backlog: a recurring, rule-shaped divergence belongs in the extractor\u002Femitter — once encoded there, DELETE the entry from the log (it no longer needs replay). Write the log with ",{"type":51,"tag":83,"props":1097,"children":1099},{"className":1098},[],[1100],{"type":57,"value":466},{"type":57,"value":1102}," semantics: re-polishing the same band×axes UPDATES its entry; a new band\u002Faxis APPENDS.",{"type":51,"tag":52,"props":1104,"children":1106},{"id":1105},"phase-0-replay-the-parity-log-if-it-exists",[1107],{"type":57,"value":1108},"PHASE 0 — Replay the parity log (if it exists)",{"type":51,"tag":60,"props":1110,"children":1111},{},[1112,1114,1119,1121,1126,1128,1133,1135,1140],{"type":57,"value":1113},"If ",{"type":51,"tag":83,"props":1115,"children":1117},{"className":1116},[],[1118],{"type":57,"value":425},{"type":57,"value":1120}," is present, run BEFORE the fresh assessment: for each ",{"type":51,"tag":83,"props":1122,"children":1124},{"className":1123},[],[1125],{"type":57,"value":473},{"type":57,"value":1127}," entry (applied\u002Fresidual, in band order), apply its ",{"type":51,"tag":83,"props":1129,"children":1131},{"className":1130},[],[1132],{"type":57,"value":784},{"type":57,"value":1134}," to the CURRENT build (the band may have just been regenerated by a reconstruct), then VERIFY by cropping that band from the source screenshot + the built page and reading BOTH — never assume the replay worked. Update the entry's ",{"type":51,"tag":83,"props":1136,"children":1138},{"className":1137},[],[1139],{"type":57,"value":879},{"type":57,"value":1141}," from what you see. Replay re-grounds on the source each time, so it survives structural drift. Then continue to Phase 1, which catches anything the log didn't cover.",{"type":51,"tag":52,"props":1143,"children":1145},{"id":1144},"phase-1-batch-assessment-always-do-this-first-one-pass",[1146],{"type":57,"value":1147},"PHASE 1 — Batch assessment (ALWAYS do this first; one pass)",{"type":51,"tag":60,"props":1149,"children":1150},{},[1151,1156,1158,1164,1166,1172,1173,1179,1181,1187,1189,1195,1197,1203,1205,1211,1213,1219,1221,1227,1229,1234],{"type":51,"tag":66,"props":1152,"children":1153},{},[1154],{"type":57,"value":1155},"Parity gate (run FIRST, before any per-section work):",{"type":57,"value":1157}," read ",{"type":51,"tag":83,"props":1159,"children":1161},{"className":1160},[],[1162],{"type":57,"value":1163},"\u003CreplicaShotsDir>\u002Fcomparison.json",{"type":57,"value":1165}," (version 2). For each page: if every scored viewport has ",{"type":51,"tag":83,"props":1167,"children":1169},{"className":1168},[],[1170],{"type":57,"value":1171},"status:\"ok\"",{"type":57,"value":106},{"type":51,"tag":83,"props":1174,"children":1176},{"className":1175},[],[1177],{"type":57,"value":1178},"score >= 0.995",{"type":57,"value":1180}," (PARITY_GATE_SCORE), AND ",{"type":51,"tag":83,"props":1182,"children":1184},{"className":1183},[],[1185],{"type":57,"value":1186},"heightMismatchRatio \u003C= 0.02",{"type":57,"value":1188}," (HEIGHT_MISMATCH_THRESHOLD), the page is ALREADY AT PARITY — skip it entirely (no section assessment, no match-section dispatch) and log ",{"type":51,"tag":83,"props":1190,"children":1192},{"className":1191},[],[1193],{"type":57,"value":1194},"gate-skip: \u003Cslug> score=\u003Cs> hmr=\u003Cr>",{"type":57,"value":1196},". A high score with ",{"type":51,"tag":83,"props":1198,"children":1200},{"className":1199},[],[1201],{"type":57,"value":1202},"heightMismatchRatio > 0.02",{"type":57,"value":1204}," does NOT pass — that combination is the short-capture artifact; treat the height mismatch itself as a high-severity finding and proceed with assessment. When the file is version 1 (no mismatch fields), the gate is score-only — never assume missing fields mean zero mismatch. If comparison.json is absent (compare never ran), there is no gate — assess every page. When ",{"type":51,"tag":83,"props":1206,"children":1208},{"className":1207},[],[1209],{"type":57,"value":1210},"fullPageScore",{"type":57,"value":1212}," is present AND the heights match (",{"type":51,"tag":83,"props":1214,"children":1216},{"className":1215},[],[1217],{"type":57,"value":1218},"originHeight === replicaHeight",{"type":57,"value":1220},"), prefer it over the crop score for the gate decision (same 0.995 bar) — it sees below the fold. When the heights differ, gate on the crop score + ",{"type":51,"tag":83,"props":1222,"children":1224},{"className":1223},[],[1225],{"type":57,"value":1226},"heightMismatchRatio",{"type":57,"value":1228}," as above: ",{"type":51,"tag":83,"props":1230,"children":1232},{"className":1231},[],[1233],{"type":57,"value":1210},{"type":57,"value":1235}," counts the magenta padding as diff by construction, so even a benign sub-threshold height delta (the 32px admin-bar artifact ≈ 0.6% on a long page) caps it below 0.995 and would make the gate-skip permanently unreachable.",{"type":51,"tag":1237,"props":1238,"children":1239},"ol",{},[1240,1261,1329,1342],{"type":51,"tag":297,"props":1241,"children":1242},{},[1243,1245,1251,1253,1259],{"type":57,"value":1244},"Read the source screenshot width: ",{"type":51,"tag":83,"props":1246,"children":1248},{"className":1247},[],[1249],{"type":57,"value":1250},"identify -format \"%w\" outputDir\u002Fscreenshots\u002Fdesktop\u002F\u003Cslug>.png",{"type":57,"value":1252},". Render the built page ONCE at THAT width: ",{"type":51,"tag":83,"props":1254,"children":1256},{"className":1255},[],[1257],{"type":57,"value":1258},"npx tsx scripts\u002F_shot.ts \"\u003CpreviewUrl>\u002F\u003Cslug>\u002F?v=1\" outputDir\u002Fscreenshots\u002Fbuilt-\u003Cslug>.png \u003CsrcWidth>",{"type":57,"value":1260}," (it scrolls to settle lazy images; do not use the Playwright MCP screenshot — it times out on tall pages).",{"type":51,"tag":297,"props":1262,"children":1263},{},[1264,1266,1272,1274,1280,1282,1288,1290,1296,1297,1303,1305,1311,1313,1319,1321,1327],{"type":57,"value":1265},"Get the built band boundaries: ",{"type":51,"tag":83,"props":1267,"children":1269},{"className":1268},[],[1270],{"type":57,"value":1271},"browser_navigate",{"type":57,"value":1273}," at ",{"type":51,"tag":83,"props":1275,"children":1277},{"className":1276},[],[1278],{"type":57,"value":1279},"\u003CsrcWidth>×900",{"type":57,"value":1281},", then ",{"type":51,"tag":83,"props":1283,"children":1285},{"className":1284},[],[1286],{"type":57,"value":1287},"browser_evaluate",{"type":57,"value":1289}," to list the band container's children (the ",{"type":51,"tag":83,"props":1291,"children":1293},{"className":1292},[],[1294],{"type":57,"value":1295},".entry-content",{"type":57,"value":150},{"type":51,"tag":83,"props":1298,"children":1300},{"className":1299},[],[1301],{"type":57,"value":1302},".wp-block-post-content",{"type":57,"value":1304}," children — a MIX of ",{"type":51,"tag":83,"props":1306,"children":1308},{"className":1307},[],[1309],{"type":57,"value":1310},"\u003Cdiv class=\"wp-block-cover\">",{"type":57,"value":1312}," heroes and ",{"type":51,"tag":83,"props":1314,"children":1316},{"className":1315},[],[1317],{"type":57,"value":1318},"\u003Csection class=\"wp-block-group\">",{"type":57,"value":1320}," bands) with each one's ",{"type":51,"tag":83,"props":1322,"children":1324},{"className":1323},[],[1325],{"type":57,"value":1326},"getBoundingClientRect()",{"type":57,"value":1328}," top + height. That gives N built bands in document order.",{"type":51,"tag":297,"props":1330,"children":1331},{},[1332,1334,1340],{"type":57,"value":1333},"For EACH band, crop the SAME vertical range from BOTH the source screenshot and the built screenshot (",{"type":51,"tag":83,"props":1335,"children":1337},{"className":1336},[],[1338],{"type":57,"value":1339},"magick \u003Cimg> -crop \u003CW>x\u003CH>+0+\u003Ctop> +repage out.png",{"type":57,"value":1341},"; source band tops scale from the built tops by the page-height ratio, or detect bg-color band boundaries) and READ the two crops side by side. Flag the band DIVERGENT on the axes you can SEE: background color, heading size\u002Fweight relative to the band, body text size, alignment (centered vs left), inter-element spacing, full-bleed vs boxed, dropped media\u002Fcontent, button style. This is a visual judgement per band, not a number diff.",{"type":51,"tag":297,"props":1343,"children":1344},{},[1345,1350],{"type":51,"tag":66,"props":1346,"children":1347},{},[1348],{"type":57,"value":1349},"Print the per-band verdict table",{"type":57,"value":1351}," (band → divergent axes, or DONE). Bands that look identical are DONE — do not touch them.",{"type":51,"tag":60,"props":1353,"children":1354},{},[1355],{"type":57,"value":1356},"This whole phase is ONE render + N band crops. It replaces eyeballing the whole page and guessing, and it does NOT chase captured @1440 numbers.",{"type":51,"tag":60,"props":1358,"children":1359},{},[1360,1365,1367,1373,1375,1381,1383,1388],{"type":51,"tag":66,"props":1361,"children":1362},{},[1363],{"type":57,"value":1364},"Diff-image hygiene:",{"type":57,"value":1366}," a ",{"type":51,"tag":83,"props":1368,"children":1370},{"className":1369},[],[1371],{"type":57,"value":1372},".diff.png",{"type":57,"value":1374}," \u002F ",{"type":51,"tag":83,"props":1376,"children":1378},{"className":1377},[],[1379],{"type":57,"value":1380},".padded.png",{"type":57,"value":1382}," marks WHERE pixels differ — never read a color, font, or size from it. Red is the difference marker; magenta (in ",{"type":51,"tag":83,"props":1384,"children":1386},{"className":1385},[],[1387],{"type":57,"value":1380},{"type":57,"value":1389},") means one side has no content at that location (layout overrun — high severity). Read every actual value from the SOURCE screenshot.",{"type":51,"tag":52,"props":1391,"children":1393},{"id":1392},"phase-2-iterate-divergent-sections-worst-first",[1394],{"type":57,"value":1395},"PHASE 2 — Iterate divergent sections (worst-first)",{"type":51,"tag":60,"props":1397,"children":1398},{},[1399],{"type":57,"value":1400},"For each divergent section, ordered by composite descending:",{"type":51,"tag":1237,"props":1402,"children":1403},{},[1404,1435,1445,1485],{"type":51,"tag":297,"props":1405,"children":1406},{},[1407,1409,1418,1420,1426,1428,1433],{"type":57,"value":1408},"Run ",{"type":51,"tag":66,"props":1410,"children":1411},{},[1412],{"type":51,"tag":83,"props":1413,"children":1415},{"className":1414},[],[1416],{"type":57,"value":1417},"match-section",{"type":57,"value":1419}," for that section index (dispatch a subagent pointed at ",{"type":51,"tag":83,"props":1421,"children":1423},{"className":1422},[],[1424],{"type":57,"value":1425},"skills\u002Fmatch-section\u002FSKILL.md",{"type":57,"value":1427},", OR apply inline if you are already that executor): apply the captured values for the divergent axes (from the spec \u002F ",{"type":51,"tag":83,"props":1429,"children":1431},{"className":1430},[],[1432],{"type":57,"value":273},{"type":57,"value":1434},") as canonicalization-safe core block attributes, render, and eyes-on compare source-crop vs built-crop until visually identical.",{"type":51,"tag":297,"props":1436,"children":1437},{},[1438,1443],{"type":51,"tag":66,"props":1439,"children":1440},{},[1441],{"type":57,"value":1442},"Re-measure ONLY that section",{"type":57,"value":1444}," (re-run the Phase-1 measurement for its index) to confirm its divergent cells cleared. Do not re-render\u002Fre-measure the whole page each time.",{"type":51,"tag":297,"props":1446,"children":1447},{},[1448,1453,1455,1461,1463,1468,1470,1475,1477,1483],{"type":51,"tag":66,"props":1449,"children":1450},{},[1451],{"type":57,"value":1452},"Log the fix.",{"type":57,"value":1454}," Once the band verifies as matched (or residual), upsert a parity-log entry for it — ",{"type":51,"tag":83,"props":1456,"children":1458},{"className":1457},[],[1459],{"type":57,"value":1460},"{band, bandIndex, divergence, sourceTarget, fix, axes, status, evidence}",{"type":57,"value":1462}," — and, if the divergence is a recurring rule-shaped gap (e.g. a capture\u002Frender defect any site would hit), set ",{"type":51,"tag":83,"props":1464,"children":1466},{"className":1465},[],[1467],{"type":57,"value":917},{"type":57,"value":1469}," with how to encode it. Save ",{"type":51,"tag":83,"props":1471,"children":1473},{"className":1472},[],[1474],{"type":57,"value":425},{"type":57,"value":1476}," after each band so a crash\u002Finterrupt never loses the record. Also save the source + built evidence crops under ",{"type":51,"tag":83,"props":1478,"children":1480},{"className":1479},[],[1481],{"type":57,"value":1482},"\u003CoutputDir>\u002Fscreenshots\u002Fparity\u002F",{"type":57,"value":1484},".",{"type":51,"tag":297,"props":1486,"children":1487},{},[1488,1493,1495,1500],{"type":51,"tag":66,"props":1489,"children":1490},{},[1491],{"type":57,"value":1492},"Bounds (anti-runaway):",{"type":57,"value":1494}," at most 3 fix cycles per band; if a band still diverges after 3, log it ",{"type":51,"tag":83,"props":1496,"children":1498},{"className":1497},[],[1499],{"type":57,"value":1078},{"type":57,"value":1501}," (with the remaining axes + why) and move on — do NOT keep grinding one band. Cap the whole page at a sane total (e.g. ≤ 2× the band count of cycles); if hit, stop and report.",{"type":51,"tag":60,"props":1503,"children":1504},{},[1505,1507,1512,1514,1520,1522,1528],{"type":57,"value":1506},"Mechanics for applying + rendering are in ",{"type":51,"tag":83,"props":1508,"children":1510},{"className":1509},[],[1511],{"type":57,"value":1425},{"type":57,"value":1513}," (BACK UP post_content to a file first; post_content via ",{"type":51,"tag":83,"props":1515,"children":1517},{"className":1516},[],[1518],{"type":57,"value":1519},"studio wp post update",{"type":57,"value":1521},"; EXACT unique-string edits, never a greedy regex; mirror into both theme pattern copies; ",{"type":51,"tag":83,"props":1523,"children":1525},{"className":1524},[],[1526],{"type":57,"value":1527},"cache flush",{"type":57,"value":1529},"). Preserve ALL content; never touch the nav template, footer part, or a band already marked DONE.",{"type":51,"tag":52,"props":1531,"children":1533},{"id":1532},"output",[1534],{"type":57,"value":1535},"Output",{"type":51,"tag":293,"props":1537,"children":1538},{},[1539,1544,1549,1554,1559,1572,1589],{"type":51,"tag":297,"props":1540,"children":1541},{},[1542],{"type":57,"value":1543},"The Phase-1 verdict table (before): every band, divergent axes flagged.",{"type":51,"tag":297,"props":1545,"children":1546},{},[1547],{"type":57,"value":1548},"Per divergent band: the axes fixed (before→after) and final MATCHED\u002FRESIDUAL.",{"type":51,"tag":297,"props":1550,"children":1551},{},[1552],{"type":57,"value":1553},"The verdict table re-run at the end (after), confirmed by looking at the crops.",{"type":51,"tag":297,"props":1555,"children":1556},{},[1557],{"type":57,"value":1558},"Total cycles spent (so a runaway is visible).",{"type":51,"tag":297,"props":1560,"children":1561},{},[1562,1570],{"type":51,"tag":66,"props":1563,"children":1564},{},[1565],{"type":51,"tag":83,"props":1566,"children":1568},{"className":1567},[],[1569],{"type":57,"value":425},{"type":57,"value":1571}," written with one entry per fix (the replayable record).",{"type":51,"tag":297,"props":1573,"children":1574},{},[1575,1580,1582,1587],{"type":51,"tag":66,"props":1576,"children":1577},{},[1578],{"type":57,"value":1579},"Promote-to-emitter backlog:",{"type":57,"value":1581}," the list of ",{"type":51,"tag":83,"props":1583,"children":1585},{"className":1584},[],[1586],{"type":57,"value":917},{"type":57,"value":1588}," notes — the recurring, rule-shaped divergences that should graduate into the extractor\u002Femitter (and then leave the log).",{"type":51,"tag":297,"props":1590,"children":1591},{},[1592,1594,1600,1602],{"type":57,"value":1593},"Before declaring done any page where at least one section was dispatched to match-section: ",{"type":51,"tag":83,"props":1595,"children":1597},{"className":1596},[],[1598],{"type":57,"value":1599},"liberate_refine_report {outputDir, slug}",{"type":57,"value":1601}," must pass for that page. A failed coverage check reopens the page — dispatch match-section to fix the accounting. Pages with zero match-section dispatches (gate-skipped, or all bands matched at assessment) have no refine\u002F",{"type":51,"tag":120,"props":1603,"children":1604},{},[1605],{"type":57,"value":1606},"\u002F directory and are exempt — do NOT call the tool for them.",{"type":51,"tag":60,"props":1608,"children":1609},{},[1610],{"type":57,"value":1611},"Never report the page MATCHED without having cropped source+built per band and looked at both. The parity log is what makes this sustainable: a future reconstruct (done for its own reason) regenerates the deterministic baseline, Phase 0 replays the log onto it, and anything that has since been encoded into the emitter is simply dropped from the log.",{"type":51,"tag":1613,"props":1614,"children":1615},"style",{},[1616],{"type":57,"value":1617},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":1619,"total":1728},[1620,1637,1654,1669,1682,1701,1713],{"slug":1621,"name":1621,"fn":1622,"description":1623,"org":1624,"tags":1625,"stars":25,"repoUrl":26,"updatedAt":1636},"adapt","build platform adapters for content extraction","Build a new platform adapter to extract content from an unsupported platform (Blogger, Ghost, Tumblr, etc.)",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1626,1629,1632,1635],{"name":1627,"slug":1628,"type":15},"Automation","automation",{"name":1630,"slug":1631,"type":15},"Data Cleaning","data-cleaning",{"name":1633,"slug":1634,"type":15},"Data Engineering","data-engineering",{"name":17,"slug":18,"type":15},"2026-05-09T05:32:13.987972",{"slug":1638,"name":1638,"fn":1639,"description":1640,"org":1641,"tags":1642,"stars":25,"repoUrl":26,"updatedAt":1653},"compose-page-blocks","compose WordPress block markup from HTML","Compose a single liberated page's WordPress block-editor markup from its rendered HTML and screenshot. Inputs are a sanitized HTML file, a desktop screenshot, the design-foundation tokens, the URL's archetype (page\u002Fpost\u002Fproduct\u002Fetc.), and the source URL. Output is a string of valid block markup that round-trips through parse_blocks, uses theme tokens (no inlined hex colors), and contains only text drawn from the source HTML. Call per-page during the streaming watch loop after extraction has produced HTML+screenshot for that URL. Use when a freshly-imported page needs `post_content` upgraded from raw HTML into block-editor markup so the replica theme's tokens, gradients, and patterns actually render.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1643,1646,1649,1652],{"name":1644,"slug":1645,"type":15},"Block Editor","block-editor",{"name":1647,"slug":1648,"type":15},"Content Creation","content-creation",{"name":1650,"slug":1651,"type":15},"HTML","html",{"name":17,"slug":18,"type":15},"2026-06-08T08:17:34.212397",{"slug":1655,"name":1655,"fn":1656,"description":1657,"org":1658,"tags":1659,"stars":25,"repoUrl":26,"updatedAt":1668},"creating-blocks","create new WordPress blocks","Templates and guidelines for creating new WordPress blocks from scratch — load this before generating block files",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1660,1661,1664,1667],{"name":1644,"slug":1645,"type":15},{"name":1662,"slug":1663,"type":15},"Full Site Editing","full-site-editing",{"name":1665,"slug":1666,"type":15},"Plugin Development","plugin-development",{"name":17,"slug":18,"type":15},"2026-06-08T08:17:49.413995",{"slug":1670,"name":1670,"fn":1671,"description":1672,"org":1673,"tags":1674,"stars":25,"repoUrl":26,"updatedAt":1681},"creating-themes","create WordPress block themes","Guidelines for creating new WordPress block themes from scratch — load this before generating theme files",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1675,1676,1677,1680],{"name":1644,"slug":1645,"type":15},{"name":1662,"slug":1663,"type":15},{"name":1678,"slug":1679,"type":15},"Themes","themes",{"name":17,"slug":18,"type":15},"2026-06-08T08:17:50.666611",{"slug":1683,"name":1683,"fn":1684,"description":1685,"org":1686,"tags":1687,"stars":25,"repoUrl":26,"updatedAt":1700},"design-foundations","build design foundation JSON from sites","Build a coherent design-foundation JSON from a liberated site — semantic color\u002Ftypography\u002Fspacing roles with evidence trails. Consumes the partial scaffold produced by liberate_design_foundation_scaffold plus aggregate HTML\u002FCSS analysis and representative rendered HTML; produces a complete design-foundation matching the schema. Call after liberation, before theme generation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1688,1691,1694,1697],{"name":1689,"slug":1690,"type":15},"Data Modeling","data-modeling",{"name":1692,"slug":1693,"type":15},"Design","design",{"name":1695,"slug":1696,"type":15},"Design System","design-system",{"name":1698,"slug":1699,"type":15},"Frontend","frontend","2026-06-08T08:17:44.457834",{"slug":1702,"name":1702,"fn":1703,"description":1704,"org":1705,"tags":1706,"stars":25,"repoUrl":26,"updatedAt":1712},"design-qa","run visual QA on WordPress themes","Visual-QA loop run after replica theme install and content import. Captures replica screenshots, applies a hard responsiveness gate at 390px AND a hard per-section visual-parity gate (measured SectionParity records, verdict computed by buildRunReport), runs qualitative vision review of source\u002Freplica pairs, checks accessibility, and drives a per-section escalation ladder of fixes via editing-themes\u002Fediting-blocks\u002Frebuild-section (R1 CSS → R2 spec-rebuild → R3 re-extract → R4a AI canonical-block rebuild → R4b deterministic styled-island floor) — escalating unresolved divergences to the operator rather than shipping them. Orchestration-internal — invoked by the replicate-with-blocks\u002Fliberate orchestrators, not directly by users.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1707,1708,1709,1710,1711],{"name":13,"slug":14,"type":15},{"name":20,"slug":21,"type":15},{"name":23,"slug":24,"type":15},{"name":1678,"slug":1679,"type":15},{"name":17,"slug":18,"type":15},"2026-06-08T08:17:45.698278",{"slug":1714,"name":1714,"fn":1715,"description":1716,"org":1717,"tags":1718,"stars":25,"repoUrl":26,"updatedAt":1727},"diagnose","debug failed data extraction processes","Debug failed or low-quality extractions by analyzing logs, probing the source site, and identifying root causes",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1719,1720,1723,1726],{"name":1633,"slug":1634,"type":15},{"name":1721,"slug":1722,"type":15},"Debugging","debugging",{"name":1724,"slug":1725,"type":15},"Logs","logs",{"name":13,"slug":14,"type":15},"2026-05-10T05:48:11.267642",21,{"items":1730,"total":1900},[1731,1750,1763,1777,1794,1809,1819,1834,1849,1860,1873,1889],{"slug":1732,"name":1732,"fn":1733,"description":1734,"org":1735,"tags":1736,"stars":1747,"repoUrl":1748,"updatedAt":1749},"annotate","collect visual feedback with browser annotation tools","Open a browser with visual annotation tools. The user clicks elements on their site and leaves feedback — the agent reads annotations and makes changes. Use this when the user wants to point at specific elements to fix, tweak, or redesign.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1737,1738,1741,1744],{"name":1698,"slug":1699,"type":15},{"name":1739,"slug":1740,"type":15},"Productivity","productivity",{"name":1742,"slug":1743,"type":15},"UX Copy","ux-copy",{"name":1745,"slug":1746,"type":15},"UX Design","ux-design",484,"https:\u002F\u002Fgithub.com\u002FAutomattic\u002Fstudio","2026-05-06T05:40:01.516544",{"slug":1751,"name":1751,"fn":1752,"description":1753,"org":1754,"tags":1755,"stars":1747,"repoUrl":1748,"updatedAt":1762},"block-content","write editable WordPress block markup","Write editable WordPress block markup for local Studio sites, including core\u002Fhtml limits, block-theme layout rules, full-width sections, validation, and skeleton-first page\u002FCSS recipes.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1756,1757,1760,1761],{"name":1644,"slug":1645,"type":15},{"name":1758,"slug":1759,"type":15},"CSS","css",{"name":1650,"slug":1651,"type":15},{"name":17,"slug":18,"type":15},"2026-05-27T07:01:55.629681",{"slug":1764,"name":1764,"fn":1765,"description":1766,"org":1767,"tags":1768,"stars":1747,"repoUrl":1748,"updatedAt":1776},"hosting-plans-helper","provide WordPress.com hosting plan information","Answer WordPress.com plan, pricing, upgrade, and feature-tier questions (plan names, what each tier unlocks — plugins, themes, custom code, SSH, hosting — and current prices) from authoritative live data. Load before answering ANY plan, pricing, or feature-gating question; never answer these from memory.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1769,1772,1775],{"name":1770,"slug":1771,"type":15},"Pricing","pricing",{"name":1773,"slug":1774,"type":15},"Reference","reference",{"name":17,"slug":18,"type":15},"2026-07-02T07:42:33.654791",{"slug":1778,"name":1778,"fn":1779,"description":1780,"org":1781,"tags":1782,"stars":1747,"repoUrl":1748,"updatedAt":1793},"liberate","migrate websites to WordPress","Import and rebuild a website from a closed platform (Wix, Squarespace, Webflow, Shopify, GoDaddy, Hostinger, HubSpot, Weebly) into a Studio WordPress site. Extracts pages\u002Fposts\u002Fproducts + media, then reconstructs the design as editable blocks + WooCommerce OR as a high-fidelity replica theme. Invoke when the user wants to migrate, import, liberate, or rebuild a site from one of these platforms.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1783,1786,1789,1792],{"name":1784,"slug":1785,"type":15},"CMS","cms",{"name":1787,"slug":1788,"type":15},"Migration","migration",{"name":1790,"slug":1791,"type":15},"Web Development","web-development",{"name":17,"slug":18,"type":15},"2026-07-09T06:47:33.454311",{"slug":1795,"name":1795,"fn":1796,"description":1797,"org":1798,"tags":1799,"stars":1747,"repoUrl":1748,"updatedAt":1808},"need-for-speed","run frontend performance audits for WordPress sites","Run a frontend performance audit on a WordPress site and get actionable optimization recommendations.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1800,1803,1804,1807],{"name":1801,"slug":1802,"type":15},"Audit","audit",{"name":1698,"slug":1699,"type":15},{"name":1805,"slug":1806,"type":15},"Performance","performance",{"name":17,"slug":18,"type":15},"2026-05-06T05:40:06.433267",{"slug":1810,"name":1810,"fn":1811,"description":1812,"org":1813,"tags":1814,"stars":1747,"repoUrl":1748,"updatedAt":1818},"plugin-recommendations","recommend WordPress plugins for site features","Choose recommended plugins and plugin-provided blocks for features core WordPress blocks do not cover - ecommerce (WooCommerce), forms and newsletters (Jetpack), online courses and quizzes (Sensei LMS), polls, surveys and ratings (Crowdsignal), spam protection (Akismet) - while keeping generated content editable and avoiding raw HTML fallbacks. Any request to sell products or build a shop, store, or storefront requires WooCommerce with products.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1815,1816,1817],{"name":1647,"slug":1648,"type":15},{"name":1665,"slug":1666,"type":15},{"name":17,"slug":18,"type":15},"2026-05-27T07:01:58.249105",{"slug":1820,"name":1820,"fn":1821,"description":1822,"org":1823,"tags":1824,"stars":1747,"repoUrl":1748,"updatedAt":1833},"rank-me-up","run on-page SEO audits for WordPress sites","Run an on-page SEO audit on a WordPress site and get actionable recommendations to improve search visibility.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1825,1826,1829,1832],{"name":1801,"slug":1802,"type":15},{"name":1827,"slug":1828,"type":15},"Marketing","marketing",{"name":1830,"slug":1831,"type":15},"SEO","seo",{"name":17,"slug":18,"type":15},"2026-05-06T05:40:05.196367",{"slug":1835,"name":1835,"fn":1836,"description":1837,"org":1838,"tags":1839,"stars":1747,"repoUrl":1748,"updatedAt":1848},"site-spec","gather specifications for new WordPress sites","Gather the site name and layout preference before building a WordPress site. Run this before creating any new site.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1840,1841,1844,1847],{"name":1692,"slug":1693,"type":15},{"name":1842,"slug":1843,"type":15},"Product Management","product-management",{"name":1845,"slug":1846,"type":15},"Specs","specs",{"name":17,"slug":18,"type":15},"2026-05-06T05:40:02.739409",{"slug":1850,"name":1850,"fn":1851,"description":1852,"org":1853,"tags":1854,"stars":1747,"repoUrl":1748,"updatedAt":1859},"studio-cli","manage local WordPress sites with Studio CLI","Use the Studio CLI to manage local WordPress sites, authentication, and preview sites. Invoke this skill when you need to run Studio CLI commands, manage sites, or troubleshoot site issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1855,1858],{"name":1856,"slug":1857,"type":15},"CLI","cli",{"name":17,"slug":18,"type":15},"2026-04-06T18:02:57.150231",{"slug":1861,"name":1861,"fn":1862,"description":1863,"org":1864,"tags":1865,"stars":1747,"repoUrl":1748,"updatedAt":1872},"taxonomist","optimize WordPress category taxonomy","Analyze and optimize a WordPress site's category taxonomy. Exports all posts, uses AI to suggest an improved category structure — merging duplicates, retiring dead categories, creating missing ones, writing descriptions, and re-categorizing posts. Run this when the user wants to clean up or improve their categories.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1866,1869,1870,1871],{"name":1867,"slug":1868,"type":15},"Content Strategy","content-strategy",{"name":1630,"slug":1631,"type":15},{"name":1830,"slug":1831,"type":15},{"name":17,"slug":18,"type":15},"2026-05-06T05:40:03.966799",{"slug":1874,"name":1874,"fn":1875,"description":1876,"org":1877,"tags":1878,"stars":1747,"repoUrl":1748,"updatedAt":1888},"visual-design","plan and execute visual design direction","Plan and execute high-quality visual direction for site creation, redesign, layout, typography, color, motion, and visual polish.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1879,1882,1883,1886],{"name":1880,"slug":1881,"type":15},"Animation","animation",{"name":1692,"slug":1693,"type":15},{"name":1884,"slug":1885,"type":15},"Typography","typography",{"name":1887,"slug":1874,"type":15},"Visual Design","2026-07-24T05:40:57.887452",{"slug":1890,"name":1890,"fn":1891,"description":1892,"org":1893,"tags":1894,"stars":1747,"repoUrl":1748,"updatedAt":1899},"visual-polish","verify and polish website visual design","Verify and polish a built or redesigned site by diagnosing rendered-DOM issues against intent and fixing them in a planned, batched screenshot-and-fix loop.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[1895,1896,1897,1898],{"name":1721,"slug":1722,"type":15},{"name":1698,"slug":1699,"type":15},{"name":20,"slug":21,"type":15},{"name":1887,"slug":1874,"type":15},"2026-06-06T07:09:59.809812",81]