Director prompt hygiene: impossible tool refs, triple-stated report contract, dead no-ship branch, persona coupling

src/agent/directors/brand-reviewer/package.ts

@@ -64,5 +64,5 @@ Missing brand sources, ambiguous scope, product-code asks.
 ## Paths
 DESIGN.md path and UI files reviewed.
 
-Never spawn. Never commit. Stay on the DESIGN.md lane.`,
+Never commit. Stay on the DESIGN.md lane.`,
 };

src/agent/directors/bruckheimer/package.ts

@@ -29,5 +29,5 @@ Read the product as a person using it: can a new user get through the first nine
 
 OUT OF LANE: implementing features, architecture sign-off, code review severity theater, fleet orchestration. Route those via Blockers to build, greybeard, critique, or skywalker.
 
-Report: Summary, Findings (product shape + discovery), Blockers, Paths.`,
+Findings: product shape and discovery, not implementation notes.`,
 };

src/agent/directors/build/package.ts

@@ -24,7 +24,6 @@ You are not a reviewer, not an orchestrator, not a doc-only planner.
 
 Before substantial repo work: follow style and philosophy conventions (baked; use_skill is not mounted on workers).
 Follow AGENTS.md and /docs. Touch only what the brief requires.
-Do not spawn sub-agents.
 
 DONE GATE: Stop when every success_criteria item from the brief is met OR explicitly blocked under Blockers. Do not invent architecture or expand the brief after criteria are satisfied.
 
@@ -34,7 +33,5 @@ REPORT MAP: Findings must map each success_criteria item → pass | fail | block
 
 API CONTRACT: Preserve existing public API sync/async and return shapes unless the brief explicitly changes them. If the brief or existing code shows a synchronous function returning a plain value (e.g. { status, body }), keep it sync — do not return a Promise / make it async just to use Web Crypto. Prefer sync libraries (node:crypto createHmac, etc.) when the public surface is sync. When the brief states a signature, match parameter order, optionality, and return type exactly. Do not change call sites to await unless the brief requires an async API.
 
-OUT OF LANE: pure exploration maps, architecture essays without code, review-only verdicts, mechanical command lists without implementing.
-
-Report: Summary, Findings, Blockers, Paths.`,
+OUT OF LANE: pure exploration maps, architecture essays without code, review-only verdicts, mechanical command lists without implementing.`,
 };

src/agent/directors/critique/package.ts

@@ -43,7 +43,7 @@ API contract check (blocking when brief specifies signatures):
 - Compare public exports against the brief and existing call sites/tests.
 - Sync → async (returning Promise when callers expect a plain value) is a blocking correctness defect.
 - Signature parameter order/optionality/return-type drift vs brief is blocking.
-- Prefer reading tests/callers; if shell is allowed, a tiny sync call that would hang on a Promise is evidence.
+- Prefer reading tests/callers; a tiny sync call via run_shell that would hang on a Promise is evidence.
 - Rank these as blocking, not style nits.
 
 Write tools are not mounted. Repro via read/shell only; recommend permanent tests for testsmith/build.
@@ -52,9 +52,5 @@ OUT OF LANE → refuse or reclassify under Blockers:
 - implementing fixes (route to build)
 - architecture portfolio without code evidence (route to greybeard)
 - visual brand / DESIGN.md (route to brand-reviewer / draper)
-- pedantic fun without evidence (route to neckbeard only if hygiene is the brief)
-
-Do not spawn. Do not apply patches. Report only.
-
-Report: Summary, Findings, Blockers, Paths.`,
+- pedantic fun without evidence (route to neckbeard only if hygiene is the brief)`,
 };

src/agent/directors/draper/package.ts

@@ -61,5 +61,5 @@ Missing references, out-of-lane asks, ambiguous scope.
 ## Paths
 Files and references inspected.
 
-Never write/edit/delete product files. Never spawn. Never commit.`,
+Never commit.`,
 };

src/agent/directors/emil/package.ts

@@ -55,25 +55,18 @@ You are a critical eye, not the hand that solves.
 
 1. Understand scope — read the relevant UI/code before judging.
 2. Form hypotheses — which laws apply.
-3. Verify — inspect code, run existing tests/linters when practical. You cannot write temp test files (write tools denied); use read/run evidence instead.
+3. Verify — inspect code, run existing tests/linters when practical; use read/run evidence, not temp test files.
 4. Confidence: VERIFIED / HIGH / MEDIUM only.
 5. Report with law + location + evidence + severity. No implementation prescriptions.
 
 OUT OF LANE → Blockers naming: build (fixes), draper (CBS visual tokens), brand-reviewer (DESIGN.md), critique (general code review), greybeard (architecture gate).
 
-# Report
+# Report shape
 
-## Summary
-Design-engineering quality assessment; critical law violations; dominant patterns.
+Summary: design-engineering quality assessment; critical law violations; dominant patterns.
+Findings: for each, Law violated | Location | Evidence | Confidence | Severity (Critical / Major / Minor).
+Blockers: missing context, out-of-lane asks, unreadable artifacts.
+Paths: files inspected.
 
-## Findings
-For each: Law violated | Location | Evidence | Confidence | Severity (Critical / Major / Minor).
-
-## Blockers
-Missing context, out-of-lane asks, unreadable artifacts.
-
-## Paths
-Files inspected.
-
-Never write/edit/delete product files. Never spawn. Never commit. Quality over quantity — three solid findings beat fifteen speculative ones.`,
+Quality over quantity — three solid findings beat fifteen speculative ones.`,
 };

src/agent/directors/explore/package.ts

@@ -21,9 +21,7 @@ FINISH BIAS: Prefer one thorough pass then report. Expand Findings, change appro
 
 FINDINGS SHAPE: Findings must be a scannable map — key paths, symbols, call flow / ownership — not optional prose dump. Cite paths. No drive-by refactors, no feature work, no review severity theater.
 
-OUT OF LANE → report Blockers naming the right director: build, plan, critique, greybeard, intern.
-
-Report: Summary, Findings, Blockers, Paths.`,
+OUT OF LANE → report Blockers naming the right director: build, plan, critique, greybeard, intern.`,
   tools: { allow: READ_TOOLS },
   spawn: { maySpawn: false },
   nudge: { maxTurns: 35 },

src/agent/directors/gaasbot/package.ts

@@ -29,9 +29,9 @@ You do not implement product code. You do not replace Greybeard (architecture re
 
 Given findings from others (or the brief): what actually blocks a release? What ships with a note? What is filed? Ask what the team is most likely getting wrong that nobody raised. Prefer hearing "do not ship" early over a late surprise.
 
-Load philosophy when judgment trade-offs matter. Stay advice-only — no write_file/edit_file/delete_file.
+Load philosophy when judgment trade-offs matter. Stay advice-only.
 
 OUT OF LANE: implementing, architecture gate ownership, eng plan authorship as PlanDirector, merge-block theater without evidence.
 
-Report: Summary, Findings (risk/sequencing advice), Blockers, Paths.`,
+Findings: risk and sequencing advice.`,
 };

src/agent/directors/greybeard/package.ts

@@ -35,7 +35,5 @@ Focus on:
 - Misalignment between product, architecture, and implementation
 - Duplication that should be refactor/API expansion instead
 
-OUT OF LANE: shipping product code, pedantic style-only nitpicking, being a second primary orchestrator.
-
-Report: Summary, Findings, Blockers, Paths.`,
+OUT OF LANE: shipping product code, pedantic style-only nitpicking, being a second primary orchestrator.`,
 };

src/agent/directors/intern/package.ts

@@ -27,9 +27,9 @@ export const internPackage: DirectorPackage = {
 
 PRIMARY INTENT: mechanical execution only. Run exactly what the brief says. No judgment, no debugging narratives, no codebase exploration, no implementation.
 
-If anything is ambiguous, missing, or fails: STOP. Report raw command output and the blocker. Do not invent next steps. Do not spawn agents. Do not load skills unless the brief names a skill to load.
+If anything is ambiguous, missing, or fails: STOP. Report raw command output and the blocker. Do not invent next steps.
 
 You are a cheap model package — stay short.
 
-Report: Summary, Findings (commands + outputs), Blockers, Paths.`,
+Findings: commands run and their outputs.`,
 };

src/agent/directors/neckbeard/package.ts

@@ -26,9 +26,9 @@ PRIMARY INTENT: adversarial pedantic review. Surface hygiene issues, nits, and r
 
 Be pedantic on purpose: naming drift, comment rot, type escape hatches, boundary validation, off-by-ones, unicode/width/escape fiddliness, dead paths, and taste-vs-defect separation. Cite file paths and concrete snippets. Separate genuine defects from taste; label each finding.
 
-Do not apply fixes. Do not write, edit, or delete product files. Do not spawn agents. Optional skills style/philosophy may sharpen the nit lens — do not load them to rewrite the product.
+Do not apply fixes. Optional skills style/philosophy may sharpen the nit lens — do not load them to rewrite the product.
 
 OUT OF LANE → report Blockers naming the right director: build (to fix), critique (correctness defects), greybeard (architecture), plan (change plans).
 
-Report: Summary, Findings (ranked nits + evidence), Blockers, Paths.`,
+Findings: ranked nits with evidence.`,
 };

src/agent/directors/plan/package.ts

@@ -15,9 +15,9 @@ export const planPackage: DirectorPackage = {
 
 PRIMARY INTENT: author concrete engineering change plans. Do not implement product code. Do not act as architecture gate (that is Greybeard).
 
-Plans must be agent-proof: files, acceptance criteria, non-goals, risks, ordered steps. Prefer interview skill when requirements are fuzzy (ask_operator / structured questions when available).
+Plans must be agent-proof: files, acceptance criteria, non-goals, risks, ordered steps. When requirements are fuzzy, note the open questions under Blockers instead of guessing — you cannot ask the operator mid-run.
 
 OUT OF LANE: shipping the change yourself, pure code review, fleet orchestration.
 
-Report: Summary, Findings (the plan), Blockers, Paths.`,
+Findings: the plan itself.`,
 };

src/agent/directors/shakespeare/package.ts

@@ -58,9 +58,7 @@ Confirm what changed and where. Summarize consistency/gap follow-ups.
 
 Write tools are mounted with no path lock. PRIMARY INTENT is still PRODUCT/ARCHITECTURE/IMPLEMENTATION — do not implement product source code, run the fleet, or act as tester/reviewer.
 
-OUT OF LANE: shipping product features, pure code review, orchestration, treating docs as optional.
-
-Report: Summary, Findings, Blockers, Paths.`;
+OUT OF LANE: shipping product features, pure code review, orchestration, treating docs as optional.`;
 
 export const shakespearePackage: DirectorPackage = {
   id: "shakespeare",

src/agent/directors/tester/package.ts

@@ -23,13 +23,11 @@ Workflow:
 1. Identify the commands or suites the brief specifies (or project defaults when clear).
 2. Run them via shell / harness-allowed tools.
 3. Capture exit codes, key failures, and paths.
-4. Report honestly — do not patch product source to make green.
+4. Report honestly — you have no product-mutation tools, so there is no way to patch source to make green.
 
-If tests fail: document failures, suspected area, and blockers. Do not write_file/edit_file product code. Suggest a re-dispatch to build or testsmith when design gaps appear.
+If tests fail: document failures, suspected area, and blockers. Suggest a re-dispatch to build or testsmith when design gaps appear.
 
-OUT OF LANE: product Write/Edit, "just quickly" fixing, redesigning the whole suite as Testsmith's primary job, fleet orchestration.
-
-Report: Summary, Findings (commands + results), Blockers, Paths.`,
+OUT OF LANE: fixing product code, "just quickly" fixing, redesigning the whole suite as Testsmith's primary job, fleet orchestration.`,
   tools: { allow: READ_TOOLS },
   spawn: { maySpawn: false },
   nudge: { maxTurns: 40 },

src/agent/directors/testsmith/package.ts

@@ -21,17 +21,15 @@ export const testsmithPackage: DirectorPackage = {
 
 PRIMARY INTENT: design test strategy and test cases for the brief. Produce clear, agent-ready coverage plans. Do not implement product code. Do not act as the primary runtime verifier (that is Tester).
 
-Design in the report (and optional notes under tmp/ only if the brief allows). Prefer:
+Design in the report. Prefer:
 - risk-based coverage and acceptance criteria from the brief
 - unit / integration / e2e boundaries when relevant
 - concrete cases: setup, action, expected result, edge/failure modes
 - what not to test and why
 
-OUT OF LANE: product Write/Edit, fixing production code, becoming the implementer, running the full verify-and-fix loop, fleet orchestration.
+OUT OF LANE: fixing production code, becoming the implementer, running the full verify-and-fix loop, fleet orchestration.
 
-You may read and search the codebase to ground the design. You must not write product source.
-
-Report: Summary, Findings (strategy + cases), Blockers, Paths.`,
+Read and search the codebase to ground the design; you have no product-mutation tools.`,
   tools: { allow: READ_TOOLS },
   spawn: { maySpawn: false },
   nudge: { maxTurns: 40 },

src/subagent/index.test.ts

@@ -47,8 +47,6 @@ import {
 } from "./index.js";
 
 import { type } from "arktype";
-import { formatDirectorSystemPrompt } from "../agent/directors/identity.js";
-import { DIRECTOR_REGISTRY } from "../agent/directors/registry.js";
 import type {
   ReactorAction,
   ReactorCapabilities,
@@ -311,28 +309,23 @@ describe("sub-agent stop helpers", () => {
     ).toBe("complete");
   });
 
-  test("shouldRequireEvidence is armed for CritiqueDirector prompt", () => {
-    expect(
-      shouldRequireEvidence({
-        systemPromptRole: formatDirectorSystemPrompt(DIRECTOR_REGISTRY.critique),
-      }),
-    ).toBe(true);
-    expect(
-      shouldRequireEvidence({
-        systemPromptRole: DIRECTOR_REGISTRY.critique.systemPrompt,
-      }),
-    ).toBe(true);
+  test("shouldRequireEvidence is armed for the critique director id", () => {
+    expect(shouldRequireEvidence({ directorId: "critique" })).toBe(true);
   });
 
   test("shouldRequireEvidence is off for greybeard even with intent=review", () => {
     expect(
       shouldRequireEvidence({
         intent: "review",
-        systemPromptRole: formatDirectorSystemPrompt(DIRECTOR_REGISTRY.greybeard),
+        directorId: "greybeard",
       }),
     ).toBe(false);
   });
 
+  test("shouldRequireEvidence is off when no directorId is resolved", () => {
+    expect(shouldRequireEvidence({ intent: "review" })).toBe(false);
+  });
+
   test("evaluateSubAgentStop does not complete a review/critique with empty readCounts even with a full envelope", () => {
     const thrashState = {
       totalToolCalls: 1,

src/subagent/nudge-director.ts

@@ -305,8 +305,7 @@ export class SubAgentDirector extends DefaultDirector {
         stop === "no-progress" ||
         stop === "turn-budget" ||
         stop === "never-acted" ||
-        stop === "never-edited" ||
-        stop === "no-ship"
+        stop === "never-edited"
       ) {
         const checkpoint =
           stop === "no-progress"
@@ -315,9 +314,7 @@ export class SubAgentDirector extends DefaultDirector {
               ? "subagent-never-acted"
               : stop === "never-edited"
                 ? "subagent-never-edited"
-                : stop === "no-ship"
-                  ? "subagent-no-ship"
-                  : "subagent-turn-budget";
+                : "subagent-turn-budget";
         const detail =
           stop === "no-progress"
             ? `identical tool call × ${this.streak.consecutiveIdentical}`

src/subagent/run.ts

@@ -283,18 +283,15 @@ function abortReasonText(signal: AbortSignal): string | undefined {
 }
 
 /**
- * Arm requireEvidence only for CritiqueDirector. Greybeard is also
+ * Arm requireEvidence only for the critique director. Greybeard is also
  * intent=review and may spawn-only then envelope; that is not a fake
  * review — do not pull it into the empty-readCounts gate.
  */
 export function shouldRequireEvidence(input: {
   intent?: TaskIntent;
-  systemPromptRole?: string;
+  directorId?: string;
 }): boolean {
-  return (
-    typeof input.systemPromptRole === "string" &&
-    input.systemPromptRole.includes("CritiqueDirector")
-  );
+  return input.directorId === "critique";
 }
 
 // Spin up an isolated, autonomous agent loop, hand it one task, and return

src/subagent/stop-policy.ts

@@ -266,7 +266,6 @@ export type SubAgentStopReason =
   | "no-progress"
   | "never-acted"
   | "never-edited"
-  | "no-ship"
   | "report-forced"
   | "incomplete-report"
   | "incomplete-report-stop";
@@ -277,7 +276,7 @@ export type SubAgentStopReason =
  * Precedence when tools are still firing:
  * no-progress (identical fingerprints) > turn-budget (hard cap).
  * Look volume never hard-stops.
- * "report-forced", "no-ship-nudge", and "incomplete-report"
+ * "report-forced" and "incomplete-report"
  * are not competing stop reasons — they are one-shot signals telling the
  * caller to inject a wrap-up / redirect nudge and keep running; turn-budget
  * remains reachable afterward. Tool-less turns end as never-acted

src/subagent/task-tool.ts

@@ -801,6 +801,7 @@ export function createTaskTool(deps: TaskToolDeps): AgentTool {
             ...(deps.onProgress !== undefined ? { onProgress: deps.onProgress } : {}),
             ...(capabilities !== undefined ? { capabilities } : {}),
             ...(systemPromptRole !== undefined ? { systemPromptRole } : {}),
+            ...(resolvedDirectorId !== undefined ? { directorId: resolvedDirectorId } : {}),
             ...(orchestrator ? { orchestrator: true, nestedDispatch: nestedDispatch! } : {}),
             maxTurns: resolvedMaxTurns,
             ...(deps.deadlineMs !== undefined ? { deadlineMs: deps.deadlineMs } : {}),

src/subagent/types.ts

@@ -98,6 +98,8 @@ export type RunSubAgentParams = {
   onProgress?: (info: { description: string; toolName: string }) => void;
   capabilities?: CapabilityFilter;
   systemPromptRole?: string;
+  /** Resolved closed-director id (e.g. "critique") when the worker is one. Structured gate key — prefer over persona-string matching in systemPromptRole. */
+  directorId?: string;
   // When true, the assembled system prompt grants this sub-agent permission
   // to call `task` to spawn further agents (orchestrator exception to the
   // no-recursion rule). Set from AgentProfile.orchestrator at dispatch time.