@tritard/waterbrother 0.8.42 → 0.8.43

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tritard/waterbrother",
-  "version": "0.8.42",
+  "version": "0.8.43",
   "description": "Waterbrother: Grok-powered coding CLI with local tools, sessions, operator modes, and approval controls",
   "type": "module",
   "bin": {

package/src/frontend.js

@@ -53,7 +53,8 @@ const BENCHMARK_SITE_TYPE_RULES = {
     "Benchmark blog mode: do not use Tailwind CDN starter theming, Font Awesome chrome, picsum/placehold imagery, fake keyboard/search chrome, or demo-only modal article previews.",
     "Benchmark blog mode: include one real editorial image surface or one authored abstract visual composition. Do not replace visuals with text placeholders.",
     "Benchmark blog mode: do not invent named authors, editors, contributors, or interview credits unless the user explicitly asked for fictional worldbuilding.",
-    "Benchmark blog mode: prefer concrete subject matter headlines and dek copy over abstract phrases like silence, attention, memory, slowness, or restraint."
+    "Benchmark blog mode: prefer concrete subject matter headlines and dek copy over abstract phrases like silence, attention, memory, slowness, or restraint.",
+    "Benchmark blog mode: use at least two clear typographic roles and finish with a designed coda or closing gesture. Do not end on a bare year or generic footer label."
   ],
   store: [
     "Benchmark ecommerce mode: prioritize conversion architecture over editorial styling. The page should sell, not just look clean.",
@@ -218,7 +219,7 @@ const BENCHMARK_STARTERS = {
     "Section 1: restrained masthead with a non-publication title treatment.",
     "Section 2: one dominant lead piece with a strong headline and one supporting block offset beside or below it.",
     "Section 3: one compact secondary list or pair of entries, not an archive rail plus topics plus notes stack.",
-    "Section 4: minimal closing footer note only if needed.",
+    "Section 4: designed closing coda only if needed; never a bare year or default footer label.",
     "No publication shell, no archive/topic chrome, no subscribe CTA, no author bylines, no contributor credits, no demo interactions, no explanation that the page is a demo.",
     "Use one strong visual surface: a real editorial image URL or an authored abstract composition made with CSS/SVG."
   ],
@@ -232,6 +233,7 @@ const BENCHMARK_STARTERS = {
     "Section 5: compact FAQ or guarantee block if still needed.",
     "No Tailwind starter shell, no fake reviews, no fake badges, no placeholder gallery, no emoji or dashed placeholder hero, no alert-based checkout, no keyboard shortcut chrome, no demo footer copy.",
     "If you do not implement a real cart drawer or checkout flow, keep cart UI passive. Do not use alert/prompt fallbacks that explain what would happen in a real Shopify store.",
+    "If you cannot support full cart behavior, keep the buy action visually strong but behaviorally passive. Never use toasts, fake checkout success, or explanatory demo flows.",
     "The hero must merchandise the product with a believable product surface and product-specific benefit language, not generic luxury-commerce filler.",
     "Use a real product image URL or an authored SVG/CSS product silhouette with material/shadow treatment. Never use text-only image placeholders."
   ]
@@ -245,7 +247,7 @@ const BENCHMARK_STARTER_VARIANTS = {
       starter: [
         "Variant shape: severe image-text split with one offset note card and one lean secondary pair.",
         "Palette should feel mineral, paper, and shadow-driven rather than warm lifestyle-editorial.",
-        "Finish with one restrained final gesture, not a generic footer label."
+        "Finish with one restrained but intentional coda band, rule block, or caption treatment. Never end on a bare year or generic footer label."
       ]
     },
     {
@@ -254,7 +256,8 @@ const BENCHMARK_STARTER_VARIANTS = {
       starter: [
         "Variant shape: oversized white space, one dominant visual surface, one short text block, one caption-like secondary band.",
         "Treat typography like exhibition text: sparse, quiet, and deliberate.",
-        "Avoid default editorial chrome entirely; let spacing and crop do the work."
+        "Avoid default editorial chrome entirely; let spacing and crop do the work.",
+        "The closing move should feel designed, not absent. Use one quiet colophon, caption band, or spatial coda."
       ]
     },
     {
@@ -304,13 +307,15 @@ const BENCHMARK_FEWSHOT_EXAMPLES = {
       "Neutral title treatment instead of Journal/LUMEN/Dispatch.",
       "One dominant editorial image, one offset supporting block, 3-4 sections max.",
       "Concrete topic language like 'Concrete Light in Osaka' or 'Domestic Brutalism in Sao Paulo'.",
-      "No bylines, no archive chrome, no demo footer, no modal article preview JS."
+      "No bylines, no archive chrome, no demo footer, no modal article preview JS.",
+      "Finish with a designed coda, caption band, or colophon rather than a bare year."
     ],
     bad: [
       "Nav items like Essays, Conversations, Archive.",
       "Labels like Featured Essay, Latest Dispatches, Notes, From the Archive.",
       "Fake author lines, contributor credits, or footer copy explaining it is a demo.",
-      "Headline/dek filler about silence, attention, memory, slowness, or restraint."
+      "Headline/dek filler about silence, attention, memory, slowness, or restraint.",
+      "A footer that is only a year or one generic label."
     ]
   },
   store: {
@@ -319,14 +324,16 @@ const BENCHMARK_FEWSHOT_EXAMPLES = {
       "Purchase stack plus trust, objections, and compact FAQ only if needed.",
       "Copy names material, dimensions, carry/use case, warranty, or care.",
       "When the prompt is generic, stay with a neutral physical product class instead of drifting into skincare/beauty campaign pages.",
-      "No alert-based checkout, no fake review counts, no fake featured-in strip, no demo footer."
+      "No alert-based checkout, no fake review counts, no fake featured-in strip, no demo footer.",
+      "No Tailwind CDN, no Font Awesome, no picsum, no toasts, no newsletter filler."
     ],
     bad: [
       "Generic luxury filler like considered life, timeless essentials, or modern living.",
       "Beauty/skincare serum landing pages when the user did not ask for that product category.",
       "Demo cart toasts, alert-based cart fallbacks, or 'in a real Shopify store' copy.",
       "Placeholder gallery or weak hero that does not merchandise the product.",
-      "Fake proof chrome like bestseller badges, sold-this-month counters, or made-up review totals."
+      "Fake proof chrome like bestseller badges, sold-this-month counters, or made-up review totals.",
+      "Tailwind starter shells, Font Awesome icons, picsum imagery, fake press/logo rows, or newsletter signup filler."
     ]
   }
 };
@@ -342,6 +349,7 @@ const BENCHMARK_GRADERS = {
     { key: "no_demo_js", label: "grader failed: benchmark store still contains demo JS behavior", pattern: /\b(?:alert\(|prompt\(|console\.log|showToast|static demo|in a real shopify store)\b/i, weight: 4, hardBlock: true },
     { key: "no_fake_cart_fallback", label: "grader failed: benchmark store still uses fake cart or checkout fallback behavior", pattern: /\b(?:cart would open here|would go to checkout|would continue to checkout|would appear here)\b/i, weight: 4, hardBlock: true },
     { key: "no_fake_proof", label: "grader failed: benchmark store still uses fake proof or badge chrome", pattern: /\b(?:best seller|bestseller|\d[\d,]*\s+reviews|\d[\d,]*\s+\w+\s+sold\s+this\s+month|sold this month|as featured in|featured in|trusted by)\b/i, weight: 4, hardBlock: true },
+    { key: "no_template_runtime", label: "grader failed: benchmark store still relies on Tailwind CDN, Font Awesome, or placeholder imagery", pattern: /\b(?:cdn\.tailwindcss\.com|fontawesome|font-awesome|picsum\.photos|placehold\.co|placeholder\.com)\b/i, weight: 5, hardBlock: true },
     { key: "no_generic_luxury_copy", label: "grader failed: benchmark store still relies on generic luxury-commerce copy", pattern: /\b(?:objects for a considered life|considered life|timeless essentials|designed for modern living|carry less\. keep more\.)\b/i, weight: 3, hardBlock: true },
     { key: "product_specific_copy", label: "grader failed: benchmark store copy is missing product-specific merchandising detail", missing: /\b(?:material|materials|capacity|fits|carry|pocket|warranty|care|dimensions?|full-grain|canvas|leather|zip|strap)\b/i, weight: 2, hardBlock: false }
   ]
@@ -387,7 +395,8 @@ const SLOP_PATTERNS = [
   { key: "fictional_publication_brand", label: "invented publication/author scaffolding", pattern: /\b(?:by [A-Z][a-z]+ [A-Z][a-z]+|photography by [A-Z][a-z]+ [A-Z][a-z]+|journal of attention|vesper|lumen|dispatches|vol\.\s*\d+|winter 20\d{2}|autumn 20\d{2}|spring 20\d{2}|summer 20\d{2}|the quarterly|editorial\.)\b/i, weight: 3 },
   { key: "author_bylines", label: "named author or contributor byline", pattern: /\b(?:by [A-Z][A-Za-z]+(?:\s+[A-Z][A-Za-z]+)+|contributor|essay by|interview by)\b/i, weight: 3 },
   { key: "fake_social_proof", label: "fake reader count or social proof", pattern: /\b(?:\d+\s+readers joined this month|\d+\s+subscribers?|trusted by)\b/i, weight: 2 },
-  { key: "placeholder_surface", label: "placeholder surface or stand-in hero", pattern: /\b(?:placeholder|visual placeholder|product representation|hero placeholder|dashed|emoji stand-in|🧥|🎧)\b/i, weight: 3 }
+  { key: "placeholder_surface", label: "placeholder surface or stand-in hero", pattern: /\b(?:placeholder|visual placeholder|product representation|hero placeholder|dashed|emoji stand-in|🧥|🎧)\b/i, weight: 3 },
+  { key: "underdesigned_footer", label: "underdesigned closing gesture", pattern: /<footer[^>]*>\s*(?:<div[^>]*>\s*)?(?:20\d{2}|[A-Z]{2,8})\s*(?:<\/div>)?\s*<\/footer>/i, weight: 2 }
 ];
 
 function getBenchmarkSiteTypeRules(siteType, benchmarkMode) {
@@ -845,6 +854,14 @@ export async function inspectFrontendArtifacts({ cwd, promptText = "", receipt =
           flags.push("benchmark store still contains generic luxury-commerce copy");
           score += 3;
         }
+        if (/\b(?:subscribe|newsletter|join)\b/i.test(htmlText)) {
+          flags.push("benchmark store still contains newsletter or signup filler");
+          score += 2;
+        }
+        if (/\b(?:LIMITED TIME OFFER|ONLY\s+\d+\s+LEFT|ships today)\b/i.test(htmlText)) {
+          flags.push("benchmark store still contains fake urgency copy");
+          score += 3;
+        }
         if (!/\b(?:material|materials|capacity|fits|carry|pocket|warranty|care|dimensions?|full-grain|canvas|leather|zip|strap)\b/i.test(htmlText)) {
           flags.push("benchmark store still lacks product-specific merchandising detail");
           score += 2;
@@ -867,6 +884,10 @@ export async function inspectFrontendArtifacts({ cwd, promptText = "", receipt =
           flags.push("benchmark blog still contains generic editorial filler copy");
           score += 3;
         }
+        if (/<footer[^>]*>\s*(?:<div[^>]*>\s*)?(?:20\d{2}|[A-Z]{2,8})\s*(?:<\/div>)?\s*<\/footer>/i.test(htmlText)) {
+          flags.push("benchmark blog closing is underdesigned");
+          score += 2;
+        }
       }
     }
   }
@@ -896,7 +917,7 @@ export function mergeFrontendSlop(base = null, extra = null) {
 }
 
 export function shouldAutoReviseFrontend({ designReview = null, slop = null, revisionCount = 0, frontend = null } = {}) {
-  const maxPasses = frontend?.benchmarkMode && frontend?.siteType === "store" ? 5 : 3;
+  const maxPasses = frontend?.benchmarkMode && frontend?.siteType === "store" ? 6 : 3;
   if (revisionCount >= maxPasses) return false;
   if (!designReview) return false;
   if (slop?.hardBlock) return true;
@@ -908,7 +929,7 @@ export function shouldAutoReviseFrontend({ designReview = null, slop = null, rev
 export function shouldForceFrontendRebuild({ frontend = null, slop = null, revisionCount = 0 } = {}) {
   if (!frontend?.benchmarkMode) return false;
   if (!slop?.hardBlock) return false;
-  if (frontend.siteType === "store") return revisionCount < 4;
+  if (frontend.siteType === "store") return revisionCount < 5;
   return revisionCount < 2;
 }
 
@@ -928,11 +949,18 @@ export function getFrontendAcceptanceFailure({ frontend = null, slop = null, des
       flags: Array.isArray(slop?.flags) ? slop.flags : []
     };
   }
+  if (frontend.siteType === "store" && designReview?.verdict && designReview.verdict !== "strong") {
+    return {
+      siteType: frontend.siteType,
+      reason: `benchmark store acceptance failed: final design review must be strong, got ${String(designReview.verdict)}`,
+      flags: Array.isArray(slop?.flags) ? slop.flags : []
+    };
+  }
   if (!slop?.hardBlock) return null;
   const flags = Array.isArray(slop.flags) ? slop.flags : [];
   if (frontend.siteType === "store") {
     const blocking = flags.filter((flag) =>
-      /Tailwind CDN|Font Awesome|placeholder product imagery|placeholder imagery|fake proof or badge chrome|fake keyboard or shortcut chrome|demo behavior|demo page explanation copy|fake cart or checkout fallback behavior|generic luxury-commerce copy|beauty\/skincare landing-page tropes|semantic drift/i.test(flag)
+      /Tailwind CDN|Font Awesome|placeholder product imagery|placeholder imagery|fake proof or badge chrome|fake keyboard or shortcut chrome|demo behavior|demo page explanation copy|fake cart or checkout fallback behavior|generic luxury-commerce copy|beauty\/skincare landing-page tropes|semantic drift|newsletter or signup filler|fake urgency copy/i.test(flag)
     );
     if (blocking.length > 0) {
       return {
@@ -982,10 +1010,13 @@ export function buildFrontendRevisionPrompt({
     "For benchmark blog tasks, do not use generic publication-shell labels like Journal, Featured Essay, Latest Dispatches, Notes, Archive, Print Edition, Submit Work, or publication-footer framing.",
     "For benchmark blog tasks, do not use essay-magazine fallback headlines like In Praise of Slowness, On Attention, The Measure of Restraint, or similar reflective placeholders.",
     "For benchmark blog tasks, do not use Tailwind CDN starter theming, Font Awesome chrome, picsum-style placeholder imagery, fake search/shortcut chrome, subscribe/newsletter CTA chrome, or demo modal article previews.",
+    "For benchmark blog tasks, do not end on a bare year or generic footer label. Close with a designed coda, caption band, or colophon-level finishing move.",
+    "For benchmark blog tasks, do not collapse to a single-font austere page. Use a stronger headline/body distinction and one additional finishing move after the hero.",
     "For benchmark store tasks, do not leave the product image area as a labeled placeholder. Use product-shaped merchandising composition, proof, and objection-handling blocks instead.",
     "For benchmark store tasks, do not use dashed placeholder boxes, emoji stand-ins, or wireframe hero surfaces. If needed, build a graphic product silhouette or material-driven composition instead.",
     "For benchmark store tasks, do not use Tailwind CDN starter theming, Font Awesome chrome, picsum-style placeholder imagery, or fake command-palette behavior.",
     "For benchmark store tasks, do not invent review counts, bestseller labels, sold-this-month counters, or as-featured-in proof unless the user explicitly requested fictional marketing chrome.",
+    "For benchmark store tasks, do not use newsletter/signup filler, fake urgency copy, or toast/demo success flows. If cart behavior is not real, keep it passive and quiet.",
     slop?.hardBlock ? "Hard-fail benchmark patterns are still present. Rebuild the page from scratch if needed rather than preserving those patterns." : "",
     "Reduce section count if needed and push one stronger asymmetrical composition instead of a sequence of balanced blocks.",
     "Simplify the page if needed. Stronger direction with fewer elements is preferred over busier generic output.",