Overhaul Greybeard prompt

Teach Greybeard named-identity architecture judgment, limited spawn
without fake caps, and search_agents blinders instead of review-theater
scheduler language.

Closes CL-7019

Author: TheGreatAxios

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

@@ -6,14 +6,61 @@ describe("greybeardPackage", () => {
     expect(greybeardPackage.id).toBe("greybeard");
   });
 
-  test("systemPrompt is real (not Placeholder)", () => {
+  test("systemPrompt is non-empty and not a Placeholder", () => {
     expect(greybeardPackage.systemPrompt.length).toBeGreaterThan(0);
     expect(greybeardPackage.systemPrompt.startsWith("Placeholder")).toBe(false);
   });
 
-  test("systemPrompt states PRIMARY INTENT and GreybeardDirector", () => {
-    expect(greybeardPackage.systemPrompt).toMatch(/PRIMARY INTENT/i);
-    expect(greybeardPackage.systemPrompt).toContain("GreybeardDirector");
+  test("systemPrompt mentions PRIMARY INTENT", () => {
+    expect(greybeardPackage.systemPrompt).toContain("PRIMARY INTENT");
+  });
+
+  test("systemPrompt identity is Greybeard / GreybeardDirector (not job-title language)", () => {
+    const p = greybeardPackage.systemPrompt;
+    expect(p).toMatch(/GreybeardDirector \(Greybeard\)/);
+    expect(p).toMatch(/architecture judgment/i);
+    expect(p).not.toMatch(/architecture director/i);
+  });
+
+  test("systemPrompt teaches judgment for architecture approach", () => {
+    const p = greybeardPackage.systemPrompt;
+    expect(p).toContain("Judge the approach");
+    expect(p).toMatch(/constraint ownership|owns constraints/i);
+    expect(p).toMatch(/hold \/ revise \/ block|verdict/i);
+    expect(p).toMatch(/backward-compatibility|backward compatibility/i);
+  });
+
+  test("systemPrompt allows limited spawn without fake caps or scheduler language", () => {
+    const p = greybeardPackage.systemPrompt;
+    expect(p).toMatch(/intern/);
+    expect(p).toMatch(/explore/);
+    expect(p).toMatch(/critique/);
+    expect(p).toMatch(/Prefer doing the review yourself/i);
+    expect(p).toMatch(/Do not invent numeric spawn caps|not a soft ladder/i);
+    expect(p).not.toMatch(/at most \d+/i);
+    expect(p).not.toMatch(/spawn at most one/i);
+    expect(p).not.toMatch(/parallel diagnostic fleet/i);
+    expect(p).not.toMatch(/turn budget/i);
+    expect(p).not.toMatch(/parameters?:/i);
+    expect(p).not.toMatch(/fan-out/i);
+  });
+
+  test("systemPrompt has Blinders against search_agents fleet discovery", () => {
+    const p = greybeardPackage.systemPrompt;
+    expect(p).toMatch(/Blinders/i);
+    expect(p).toMatch(/do not call search_agents/i);
+    expect(p).toMatch(/even when nested/i);
+  });
+
+  test("systemPrompt guides quality without enforcement theater", () => {
+    const p = greybeardPackage.systemPrompt;
+    expect(p).toMatch(/Guide quality/i);
+    expect(p).toMatch(/enforcement theater/i);
+  });
+
+  test("systemPrompt forbids spawning build and names off-list directors", () => {
+    expect(greybeardPackage.systemPrompt).toContain("Do not spawn build");
+    expect(greybeardPackage.systemPrompt).not.toMatch(/\bspawn implement\b/);
   });
 
   test("spawn.maySpawn is true with limited allowlist", () => {
@@ -33,17 +80,6 @@ describe("greybeardPackage", () => {
     expect(allow).not.toContain("plan");
   });
 
-  test("systemPrompt forbids spawning implement and names build as off-list", () => {
-    expect(greybeardPackage.systemPrompt).not.toMatch(/\bspawn implement\b/);
-    expect(greybeardPackage.systemPrompt).toContain("Do not spawn build");
-  });
-
-  test("systemPrompt forbids parallel diagnostic fleets", () => {
-    expect(greybeardPackage.systemPrompt).toMatch(/do the review yourself/i);
-    expect(greybeardPackage.systemPrompt).toMatch(/spawn at most one intern/i);
-    expect(greybeardPackage.systemPrompt).toMatch(/never spawn a parallel diagnostic fleet/i);
-  });
-
   test("tools.allow is orchestrator surface without product writes", () => {
     const allow = greybeardPackage.tools?.allow ?? [];
     expect(allow).toContain("task");
@@ -62,7 +98,7 @@ describe("greybeardPackage", () => {
   });
 
   test("primaryIntent and outOfLane match greybeard lane", () => {
-    expect(greybeardPackage.primaryIntent).toBe("Architecture review; limited spawn");
+    expect(greybeardPackage.primaryIntent).toBe("Architecture judgment; limited spawn");
     expect(greybeardPackage.outOfLane).toContain("shipping product code");
     expect(greybeardPackage.outOfLane).toContain("pedantic style-only nitpicking");
   });

src/agent/directors/greybeard/package.ts

@@ -2,14 +2,14 @@ import type { DirectorPackage } from "../types.js";
 import { ORCHESTRATOR_TOOLS } from "../tool-sets.js";
 
 /**
- * Architecture review leaf with limited spawn (CL-5821).
- * Evidence via intern/explore/critique only — never ships product code.
+ * Greybeard nested orchestrator (CL-7019).
+ * Architecture judgment with limited spawn — never ships product code.
  */
 export const greybeardPackage: DirectorPackage = {
   id: "greybeard",
-  primaryIntent: "Architecture review; limited spawn",
+  primaryIntent: "Architecture judgment; limited spawn",
   outOfLane: ["shipping product code", "pedantic style-only nitpicking"],
-  description: "Architecture review leaf",
+  description: "Architecture judgment leaf",
   optionalSkills: ["style", "philosophy"],
   tools: { allow: ORCHESTRATOR_TOOLS },
   spawn: {
@@ -18,22 +18,26 @@ export const greybeardPackage: DirectorPackage = {
   },
   modelRole: "review",
   tier: "nested-orchestrator",
-  systemPrompt: `You are GreybeardDirector, a specialist in Corbits Code.
+  systemPrompt: `You are GreybeardDirector (Greybeard), a specialist in Corbits Code.
 
-PRIMARY INTENT: architecture review. Judge soundness, constraint ownership, and backward-compatibility implications. Do not fix or ship product code.
+PRIMARY INTENT: architecture judgment. Judge approach soundness, constraint ownership, and backward-compatibility implications. Teach what holds and what does not. Do not fix or ship product code.
 
-Load style and philosophy when reviewing plans or approaches — skills are active constraints, not background docs.
+You are Greybeard — not a second Skywalker, not Critique (code defects with evidence), not Build. Your value is architectural judgment, not legwork or implementation.
 
-You may spawn only intern, explore, and critique for evidence gathering. Do not spawn build, plan, skywalker, or other directors. Your value is analysis, not legwork or implementation.
+Judge the approach:
+1. Name the architectural claim under review (boundary, ownership, invariant, or BC surface).
+2. Decide whether the proposed approach owns constraints at the right layer — or only chases symptoms.
+3. Call out holes, anti-patterns, missing invariants, product/architecture/implementation misalignment, and duplication that should be refactor or API expansion instead.
+4. Rank risks for long-term maintainability and backward compatibility.
+5. Report a clear verdict: hold / revise / block — with the why, not a checklist theater.
 
-Do the review yourself. Spawn at most one intern, explore, or critique evidence leaf when a single unknown path blocks you. Never spawn a parallel diagnostic fleet.
+Spawn only when a concrete unknown blocks that judgment. Package spawn rules allow intern (mechanical shell), explore (map/read), and critique (code evidence). Prefer doing the review yourself with mounted read/search tools. Do not invent numeric spawn caps or act as a scheduler — width follows the unknown, not a soft ladder.
 
-Focus on:
-- Architectural holes, anti-patterns, missing invariants
-- Constraint ownership (fixed at the right layer, not symptom-chasing)
-- BC implications and long-term maintainability
-- Misalignment between product, architecture, and implementation
-- Duplication that should be refactor/API expansion instead
+Blinders: do not call search_agents to discover the fleet (even when nested). You already know the limited spawn set; stay inside it. Do not spawn build, plan, skywalker, or other directors outside the allowlist.
 
-OUT OF LANE: shipping product code, pedantic style-only nitpicking, being a second primary orchestrator.`,
+Guide quality — advise what good architecture looks like for this change. Do not assert enforcement theater (fake caps, pretend runtime gates, or "must spawn N" rules the harness does not enforce).
+
+Before substantial review work: follow style and philosophy conventions (baked; use_skill is not mounted on workers).
+
+OUT OF LANE: shipping product code, pedantic style-only nitpicking, being a second primary orchestrator, discovering or dispatching the full fleet.`,
 };