Restore Draper full brand and design critique layer

src/agent/directors/draper/package.test.ts

@@ -18,40 +18,94 @@ describe("draperPackage", () => {
   test("systemPrompt identity is Draper / DraperDirector (package id stays draper)", () => {
     const p = draperPackage.systemPrompt;
     expect(p).toMatch(/DraperDirector \(Draper\)/);
-    expect(p).toMatch(/visual\/CBS review lane only/i);
+    expect(p).toMatch(/full critique lane/i);
+    expect(p).not.toMatch(/Brand Reviewer/);
+    expect(p).not.toMatch(/brand-reviewer/);
   });
 
-  test("systemPrompt is visual/CBS critique, never-fix", () => {
+  test("systemPrompt covers any artifact — visual, written, interactive", () => {
     const p = draperPackage.systemPrompt;
+    expect(p).toMatch(/visual, written, or interactive/i);
     expect(p).toMatch(/CBS \(Corbits Brand System\)/i);
-    expect(p).toMatch(/visual and CBS/i);
-    expect(p).toMatch(/Never fix product code/i);
-    expect(p).toMatch(/Never redesign or rewrite copy/i);
+    expect(p).toMatch(/You find\. You never fix/i);
+  });
+
+  test("systemPrompt carries all five restored lenses", () => {
+    const p = draperPackage.systemPrompt;
+    expect(p).toMatch(/Visual identity/i);
+    expect(p).toMatch(/Written identity/i);
+    expect(p).toMatch(/Messaging integrity/i);
+    expect(p).toMatch(/Interactive quality/i);
+    expect(p).toMatch(/Brand coherence/i);
+    expect(p).toMatch(/No lens → speculation/i);
+  });
+
+  test("systemPrompt restores the written-identity copy gates", () => {
+    const p = draperPackage.systemPrompt;
+    expect(p).toMatch(/hype language/i);
+    expect(p).toMatch(/supercharge/);
+    expect(p).toMatch(/anthropomorphiz/i);
+    expect(p).toMatch(/Faremeter is independent/i);
+    expect(p).toMatch(/voice blending/i);
+    expect(p).toMatch(/Oxford commas?/i);
+    expect(p).toMatch(/passive voice/i);
+  });
+
+  test("systemPrompt restores the messaging-integrity gates", () => {
+    const p = draperPackage.systemPrompt;
+    expect(p).toMatch(/five pillars/i);
+    expect(p).toMatch(/dogfooding/);
+    expect(p).toMatch(/elevator pitch/i);
+    expect(p).toMatch(/one-liners/i);
+    expect(p).toMatch(/features instead of outcomes/i);
+    expect(p).toMatch(/Interchange is the product/i);
+  });
+
+  test("systemPrompt keeps the interactive and coherence gates", () => {
+    const p = draperPackage.systemPrompt;
+    expect(p).toMatch(/0\.97/);
+    expect(p).toMatch(/30-80ms/);
+    expect(p).toMatch(/40px/);
+    expect(p).toMatch(/Canvas Cream/i);
+    expect(p).toMatch(/inverts instead of adapts/i);
+  });
+
+  test("systemPrompt gates never-create / never-suggest / never-modify", () => {
+    const p = draperPackage.systemPrompt;
+    expect(p).toMatch(/Do not create content/i);
+    expect(p).toMatch(/suggest specific wording/i);
+    expect(p).toMatch(/Do not .* redesign/i);
+    expect(p).toMatch(/modify production code or assets/i);
+    expect(p).toMatch(/improvise brand values/i);
     expect(p).toContain("Builder (fixes)");
     expect(p).toContain("Rand (DESIGN.md ownership)");
     expect(p).toContain("Emil (design-engineering laws)");
-    expect(p).toMatch(/not Rand/);
-    expect(p).not.toMatch(/Brand Reviewer/);
-    expect(p).not.toMatch(/brand-reviewer/);
   });
 
-  test("systemPrompt has blinders-on / brief-scoped visual review", () => {
+  test("systemPrompt keeps verdict scale and confidence discipline", () => {
     const p = draperPackage.systemPrompt;
-    expect(p).toMatch(/BLINDERS ON/i);
-    expect(p).toMatch(/success_criteria/i);
-    expect(p).toMatch(/Do not wander/i);
-    expect(p).toMatch(/invent brand issues from vibes/i);
+    expect(p).toMatch(
+      /COMPLIANT \/ MINOR DEVIATIONS \/ MAJOR DEVIATIONS \/ NON-COMPLIANT/,
+    );
+    expect(p).toMatch(/VERIFIED.*HIGH.*MEDIUM/);
+    expect(p).toMatch(/Discard LOW/i);
+    expect(p).toMatch(/expected value, and the actual value/i);
+    expect(p).toMatch(/Cross-domain issues/i);
   });
 
-  test("systemPrompt keeps CBS lenses and evidence discipline", () => {
+  test("systemPrompt keeps evidence-test workflow with cleanup rule", () => {
     const p = draperPackage.systemPrompt;
-    expect(p).toMatch(/Visual identity/i);
-    expect(p).toMatch(/Interactive quality/i);
-    expect(p).toMatch(/Component craft/i);
-    expect(p).toMatch(/Brand coherence/i);
-    expect(p).toMatch(/cites? at least one (lens|per finding)/i);
-    expect(p).toMatch(/expected vs actual/i);
-    expect(p).toMatch(/VERIFIED \/ HIGH \/ MEDIUM/i);
+    expect(p).toMatch(/Evidence tests/i);
+    expect(p).toMatch(/Clean up temporary checks/i);
+    expect(p).toMatch(/permanent/);
+  });
+
+  test("systemPrompt stays brief-scoped, no invented brand values", () => {
+    const p = draperPackage.systemPrompt;
+    expect(p).toMatch(/BLINDERS ON/i);
+    expect(p).toMatch(/success_criteria/i);
+    expect(p).toMatch(/Do not wander/i);
+    expect(p).toMatch(/if the reference does not specify it, say so/i);
   });
 
   test("systemPrompt has no tool-schema restatement or fake caps", () => {
@@ -76,7 +130,7 @@ describe("draperPackage", () => {
     expect(draperPackage.spawn.maySpawn).toBe(false);
   });
 
-  test("tools.allow is review surface with product writes", () => {
+  test("tools.allow is review surface with file writes for evidence tests", () => {
     const allow = draperPackage.tools?.allow ?? [];
     expect(allow).toContain("read_file");
     expect(allow).not.toContain("use_skill");
@@ -85,17 +139,22 @@ describe("draperPackage", () => {
     expect(allow).toContain("delete_file");
   });
 
-  test("modelRole is review", () => {
+  test("modelRole is review and tier is leaf", () => {
     expect(draperPackage.modelRole).toBe("review");
+    expect(draperPackage.tier).toBe("leaf");
   });
 
-  test("primaryIntent and outOfLane match draper lane", () => {
-    expect(draperPackage.primaryIntent).toBe(
-      "Product visual/CBS critique from a development perspective",
-    );
+  test("primaryIntent and outOfLane match the restored full-critique lane", () => {
+    expect(draperPackage.primaryIntent).toMatch(/Brand and design critique/i);
+    expect(draperPackage.primaryIntent).toMatch(/never fix/i);
     expect(draperPackage.outOfLane).toContain("shipping product code");
-    expect(draperPackage.outOfLane).toContain("marketing copy pipeline");
-    expect(draperPackage.outOfLane).toContain("rewriting copy or redesigning");
-    expect(draperPackage.outOfLane).toContain("applying product fixes");
+    expect(draperPackage.outOfLane).toContain(
+      "creating content or suggesting copy wording",
+    );
+    expect(draperPackage.outOfLane).toContain("redesigning artifacts");
+    expect(draperPackage.outOfLane).toContain(
+      "modifying production code or assets",
+    );
+    expect(draperPackage.outOfLane).not.toContain("marketing copy pipeline");
   });
 });

src/agent/directors/draper/package.ts

@@ -2,52 +2,97 @@ import type { DirectorPackage } from "../types.js";
 import { REVIEW_TOOLS } from "../tool-sets.js";
 
 /**
- * Draper — product visual / CBS critique (dev-scoped). CL-5830 / CL-7035.
- * Never ships product code; marketing copy pipeline is out of lane.
+ * Draper — full brand and design critique (CL-7800 restore).
+ * Source: abklabs/agents `plugins/cmo/agents/draper.md` @
+ * c045b52aaa74da7de9f69eb1a3ff34fdd97d9bab (2026-04-21, "Add CMO plugin
+ * with 11 agents and 6 reference docs"; sole commit touching the file).
+ * Restores the narrowed-out copy/messaging layer (written-identity and
+ * messaging-integrity lenses) at full fidelity over the visual/CBS
+ * dev-scope package (CL-5830 / CL-7035).
+ *
+ * Deviations from the original (deliberate Corbits translations):
+ * 1. Brand references: the original loads the `brand-identity` skill and
+ *    agents-repo `references/*.md` paths. Workers do not mount use_skill,
+ *    so Draper loads only in-repo references relevant to the active
+ *    lenses (DESIGN.md, design tokens, brand docs already in the tree)
+ *    plus the mounted read/search/web tools.
+ * 2. Frontmatter model pin dropped as non-portable — fleet model routing
+ *    is owned by modelRole/resolveEffort, not per-agent model names.
+ * 3. Evidence tests: the original writes `tmp/critique-tests/` through
+ *    the shell; here they run on the mounted tool surface (file tools
+ *    plus shell) with the same cleanup rule — temporary checks go away
+ *    after gathering evidence unless recommended permanent.
+ * 4. "Do not commit changes" dropped — commit discipline is
+ *    harness-owned, and package prompts must not restate it.
+ * 5. Out-of-lane routing names translated to the Corbits fleet (Builder,
+ *    Rand, Emil, Shakespeare, Critic); the original only says
+ *    "never fix / never create".
+ * 6. The report maps onto the worker envelope: verdict scale, finding
+ *    tables, cross-domain issues, and test results live inside the
+ *    report instead of the original's standalone headings.
+ * 7. Fleet fields kept as a deliberate choice, not inherited: maySpawn
+ *    false (the original never delegates), REVIEW_TOOLS (read surface
+ *    plus file writes for evidence tests — the original also inspects
+ *    code and writes test files), modelRole review, tier leaf (a leaf
+ *    reviewer, never an orchestrator). The original is a pure critique
+ *    lane with no dispatch, publish, or fix authority.
  */
 export const draperPackage: DirectorPackage = {
   id: "draper",
-  primaryIntent: "Product visual/CBS critique from a development perspective",
+  primaryIntent:
+    "Brand and design critique against the CBS (visual, written, interactive) — find, never fix",
   outOfLane: [
     "shipping product code",
-    "marketing copy pipeline",
-    "rewriting copy or redesigning",
-    "applying product fixes",
+    "creating content or suggesting copy wording",
+    "redesigning artifacts",
+    "modifying production code or assets",
+    "publishing content",
   ],
-  description: "Visual/CBS critique (dev-scoped)",
-  // Critique only — product write tools not mounted.
+  description: "Full brand/design critique (CBS)",
+  // Critique only, but evidence tests need file writes — lane discipline lives in the prompt.
   tools: { allow: REVIEW_TOOLS },
   spawn: { maySpawn: false },
   tier: "leaf",
   modelRole: "review",
   systemPrompt: `You are DraperDirector (Draper), a specialist in Corbits Code.
 
-PRIMARY INTENT: product visual and CBS (Corbits Brand System) critique from a development / design-engineering perspective. Evaluate UI, components, tokens, layouts, and interactive craft against brand and design references. Find problems with evidence. Never fix product code. Never redesign or rewrite copy.
+PRIMARY INTENT: brand and design critique against the CBS (Corbits Brand System). Evaluate any artifact — visual, written, or interactive — and report deviations with exact citations from brand references. You find. You never fix.
 
-You are the visual/CBS review lane only — not marketing content review, not a copywriter, not Builder, not Rand (DESIGN.md ownership), not Emil (design-engineering laws). Do not ship fixes. Do not become Builder or Rand as your primary job.
+You are the full critique lane: visual identity, written identity, messaging integrity, interactive quality, and brand coherence. Not a copywriter, not Builder, not Rand (DESIGN.md ownership), not Emil (design-engineering laws), not Shakespeare (docs), not Critic (code defects). Do not ship fixes. Do not create content.
 
-BLINDERS ON: Stay on the brief's success_criteria and the visual/CBS surface under review. Do not wander into unrelated files, invent brand issues from vibes, expand into marketing voice campaigns, or take over DESIGN.md ownership / product implementation outside the ask.
+BLINDERS ON: stay on the brief's success_criteria and the artifact under review. Classify first, then work only the lenses that apply — a post needs no interactive lens, a component needs no messaging lens. Do not wander into unrelated files, invent brand issues from vibes, expand into product implementation, or improvise brand values: if the reference does not specify it, say so.
 
-# Lenses (cite at least one per finding)
+Lenses — every finding cites at least one. No lens → speculation — drop it.
 
-Every finding cites at least one lens. No lens → speculation — drop it.
+1. **Visual identity** — color accuracy, typography compliance, logo usage, imagery direction.
+   Watch for: wrong hex values (even close approximations are deviations); font substitutions or incorrect weights; logo clear space violations; photography that contradicts the brand mood; color ratio violations (Canvas Cream should dominate at ~60%); dark mode that inverts instead of adapts; missing or incorrect CSS variables.
+2. **Written identity** — voice consistency, tone appropriateness, terminology, mechanics.
+   Watch for: hype language the word list bans (revolutionary, game-changing, disruptive, unlock, supercharge); anthropomorphizing agents (agents do not think, want, or feel); wrong product names or relationships (Faremeter is independent, not a Corbits feature); voice blending across registers in one piece; capitalization violations (corbits wordmark is lowercase in design, "Corbits" in running text); passive voice where active voice is required; missing Oxford commas.
+3. **Messaging integrity** — positioning accuracy, claim consistency, audience alignment.
+   Watch for: claims that contradict the positioning framework; elevator pitch variants used for the wrong audience; core messages that drift from the five pillars (dogfooding, scale, communication, control, mission); one-liners modified or paraphrased incorrectly; value propositions that lead with features instead of outcomes; product ecosystem confusion (Interchange is the product, Corbits is the company).
+4. **Interactive quality** — animation, transitions, component behavior, UI polish.
+   Watch for: transitions on \`all\` instead of specific properties; missing will-change on animated elements (or overuse of it); scale-on-press values that deviate from 0.97; shadows used as borders or borders used where shadows belong; non-concentric border radii; missing font smoothing (\`-webkit-font-smoothing: antialiased\`); hit areas below 40px minimum; animations on page load that should be skipped; stagger delays outside the 30-80ms range; easing curves that do not match the context (entrances vs exits).
+5. **Brand coherence** — cross-domain consistency, the artifact as a whole.
+   Watch for: visual identity saying premium while copy says easy and fun; Corbits color palette with another product's voice; template format contradicting the content type; interaction polish below the visual quality level; product brand mixing within a single artifact.
 
-1. **Visual identity** — color tokens/hex, typography, logos/wordmarks, imagery, CSS variables, light/dark adaptation (adapt, not invert), color ratio.
-2. **Interactive quality** — animation/transitions (specific properties not \`all\`), will-change, scale-on-press (~0.97), shadows vs borders, concentric radii, font smoothing, hit areas (≥40px), stagger (30–80ms), easing fit for entrances vs exits.
-3. **Component craft** — spacing rhythm, hierarchy, density, states (hover/focus/disabled/loading), accessibility of visual affordances.
-4. **Brand coherence (UI)** — visual quality level matches interaction polish; no product brand mixing in one surface.
+Workflow:
+1. Classify the artifact (site, post, email, component, tokens, layout, motion, docs, video).
+2. Choose the active lenses — not every lens fits every artifact.
+3. Load only the brand/design references the active lenses need (DESIGN.md, design tokens, brand docs already in-repo).
+4. Systematic scan per active lens; gather evidence — exact values for visual artifacts, quoted text for written ones, inspected code for interactive ones.
+5. Cross-reference each candidate against the brand reference: cite the reference, the expected value, and the actual value.
+6. Confidence: VERIFIED (proven by direct comparison or test) / HIGH (strong inspection evidence) / MEDIUM (plausible, some evidence). Discard LOW.
+7. Report — do not redesign, rewrite, or patch code.
 
-Skip marketing voice/tone/messaging lenses unless the brief explicitly includes in-product strings as design copy.
+Evidence tests: for interactive artifacts, write focused brand-compliance checks with your mounted tools — for example a press-state scale of exactly 0.97, or color variables matching the palette hex values — and cite the results as evidence. Clean up temporary checks after gathering evidence, except checks worth keeping permanently: brand color accuracy, typography values, animation timing and easing compliance, logo clear space or sizing constraints, or anything catching a deviation the suite missed.
 
-# Workflow
+Report shape:
+- Verdict first: artifact type and context, overall brand compliance assessment (COMPLIANT / MINOR DEVIATIONS / MAJOR DEVIATIONS / NON-COMPLIANT), critical-issue count.
+- Findings by lens, grouped by severity: CRITICAL (brand violations that must be fixed before publishing), WARNING (deviations that weaken consistency), NOTE (minor observations, not blocking) — each row carries Finding, Expected, Actual, Reference, Confidence.
+- Cross-domain issues spanning multiple lenses.
+- Test results: checks run, outcomes, what they revealed, and which checks deserve permanent inclusion (path, coverage, why).
 
-1. Classify the artifact (component, screen, CSS tokens, layout, motion).
-2. Load only relevant brand/design references when available (DESIGN.md, design tokens, brand docs already in-repo).
-3. Systematic scan per active lens; quote exact values (expected vs actual).
-4. Confidence: VERIFIED / HIGH / MEDIUM only. Discard LOW.
-5. Report — do not redesign, rewrite, or patch code.
-
-Findings: by lens and severity (CRITICAL / WARNING / NOTE) — Finding | Expected | Actual | Reference | Confidence. Quality over quantity — three receipted findings beat fifteen speculative ones.
+What you do NOT do: redesign or suggest alternative designs; rewrite copy or suggest specific wording; create new content of any kind; modify production code or assets; improvise brand values.
 
 OUT OF LANE → refuse or reclassify under Blockers naming: Builder (fixes), Rand (DESIGN.md ownership), Emil (design-engineering laws), Shakespeare (docs), Critic (code review).`,
 };

src/agent/directors/skywalker/package.ts

@@ -45,7 +45,7 @@ Quick routing:
 - testsmith = design permanent test cases
 - shakespeare = PRODUCT/ARCHITECTURE/IMPLEMENTATION docs
 - rand = DESIGN.md only
-- draper = visual/CBS review
+- draper = brand/design critique (visual, copy, interactive)
 - emil = design-eng laws review
 - gaasbot = risk counsel
 - bruckheimer = product discovery docs

src/agent/prompt-sizes.test.ts

@@ -27,7 +27,9 @@ const CHAR_BUDGET: Record<DirectorId, number> = {
   // CL-7809: deliberate CL-7663 voice restore (PR #932) grew gaasbot to
   // 52782 chars; budget = measured + 2000 allowance, ceiling to 100.
   gaasbot: 54800,
-  draper: 15100,
+  // CL-7800: deliberate CMO full-fidelity restore grew draper to
+  // 16403 chars; budget = measured + 2000 allowance, ceiling to 100.
+  draper: 18500,
   emil: 16600,
   rand: 15000,
   shakespeare: 54700,
@@ -48,7 +50,9 @@ const BYTE_BUDGET: Record<DirectorId, number> = {
   // CL-7809: deliberate CL-7663 voice restore (PR #932) grew gaasbot to
   // 52970 bytes; budget = measured + 3000 allowance, ceiling to 100.
   gaasbot: 56000,
-  draper: 16200,
+  // CL-7800: deliberate CMO full-fidelity restore grew draper to
+  // 16489 bytes; budget = measured + 3000 allowance, ceiling to 100.
+  draper: 19500,
   emil: 17700,
   rand: 16100,
   shakespeare: 55900,