gsd-pi 2.80.0 → 2.81.0-dev.3cddbbba2

package/src/resources/extensions/gsd/tests/db-transaction.test.ts

@@ -0,0 +1,110 @@
+// Project/App: GSD-2
+// File Purpose: Tests for DB transaction depth and rollback helpers.
+
+import { describe, test } from "node:test";
+import assert from "node:assert/strict";
+import { createDbTransactionRunner, type DbTransactionControls } from "../db-transaction.ts";
+
+class FakeTransactionControls implements DbTransactionControls {
+  readonly calls: string[] = [];
+  failCall = new Set<string>();
+
+  begin(): void {
+    this.record("BEGIN");
+  }
+
+  beginRead(): void {
+    this.record("BEGIN DEFERRED");
+  }
+
+  commit(): void {
+    this.record("COMMIT");
+  }
+
+  rollback(): void {
+    this.record("ROLLBACK");
+  }
+
+  private record(call: string): void {
+    this.calls.push(call);
+    if (this.failCall.has(call)) throw new Error(`failed ${call}`);
+  }
+}
+
+describe("db-transaction", () => {
+  test("commits successful write transactions", () => {
+    const runner = createDbTransactionRunner();
+    const controls = new FakeTransactionControls();
+
+    const result = runner.transaction(controls, () => {
+      assert.equal(runner.isInTransaction(), true);
+      return "ok";
+    });
+
+    assert.equal(result, "ok");
+    assert.equal(runner.isInTransaction(), false);
+    assert.deepEqual(controls.calls, ["BEGIN", "COMMIT"]);
+  });
+
+  test("rolls back failed write transactions and clears depth", () => {
+    const runner = createDbTransactionRunner();
+    const controls = new FakeTransactionControls();
+
+    assert.throws(
+      () => runner.transaction(controls, () => {
+        throw new Error("boom");
+      }),
+      /boom/,
+    );
+
+    assert.equal(runner.isInTransaction(), false);
+    assert.deepEqual(controls.calls, ["BEGIN", "ROLLBACK"]);
+  });
+
+  test("nested transactions do not issue nested BEGIN or COMMIT", () => {
+    const runner = createDbTransactionRunner();
+    const controls = new FakeTransactionControls();
+
+    runner.transaction(controls, () => {
+      runner.transaction(controls, () => {
+        assert.equal(runner.isInTransaction(), true);
+      });
+    });
+
+    assert.deepEqual(controls.calls, ["BEGIN", "COMMIT"]);
+  });
+
+  test("failed BEGIN does not mark transaction depth active", () => {
+    const runner = createDbTransactionRunner();
+    const controls = new FakeTransactionControls();
+    controls.failCall.add("BEGIN");
+
+    assert.throws(() => runner.transaction(controls, () => undefined), /failed BEGIN/);
+
+    assert.equal(runner.isInTransaction(), false);
+    assert.deepEqual(controls.calls, ["BEGIN"]);
+  });
+
+  test("read transactions log rollback failures and clear depth", () => {
+    const runner = createDbTransactionRunner();
+    const controls = new FakeTransactionControls();
+    const rollbackErrors: Error[] = [];
+    controls.failCall.add("ROLLBACK");
+
+    assert.throws(
+      () => runner.readTransaction(
+        controls,
+        () => {
+          throw new Error("read failed");
+        },
+        (error) => rollbackErrors.push(error),
+      ),
+      /read failed/,
+    );
+
+    assert.equal(runner.isInTransaction(), false);
+    assert.deepEqual(controls.calls, ["BEGIN DEFERRED", "ROLLBACK"]);
+    assert.equal(rollbackErrors.length, 1);
+    assert.match(rollbackErrors[0].message, /failed ROLLBACK/);
+  });
+});

package/src/resources/extensions/gsd/tests/db-verification-evidence-schema.test.ts

@@ -0,0 +1,76 @@
+// Project/App: GSD-2
+// File Purpose: Tests for verification evidence schema helpers.
+
+import { describe, test } from "node:test";
+import assert from "node:assert/strict";
+import {
+  dedupeVerificationEvidenceRows,
+  ensureVerificationEvidenceDedupIndex,
+} from "../db-verification-evidence-schema.ts";
+import type { DbAdapter, DbStatement } from "../db-adapter.ts";
+
+class FakeStatement implements DbStatement {
+  private readonly row: Record<string, unknown> | undefined;
+
+  constructor(row: Record<string, unknown> | undefined) {
+    this.row = row;
+  }
+
+  run(): unknown {
+    return undefined;
+  }
+
+  get(): Record<string, unknown> | undefined {
+    return this.row;
+  }
+
+  all(): Record<string, unknown>[] {
+    return [];
+  }
+}
+
+class FakeAdapter implements DbAdapter {
+  readonly execCalls: string[] = [];
+  hasDedupIndex = false;
+
+  exec(sql: string): void {
+    this.execCalls.push(sql);
+  }
+
+  prepare(): DbStatement {
+    return new FakeStatement(this.hasDedupIndex ? { present: 1 } : undefined);
+  }
+
+  close(): void {}
+}
+
+describe("db-verification-evidence-schema", () => {
+  test("dedupeVerificationEvidenceRows keeps the first row for each evidence identity", () => {
+    const db = new FakeAdapter();
+
+    dedupeVerificationEvidenceRows(db);
+
+    assert.equal(db.execCalls.length, 1);
+    assert.match(db.execCalls[0], /DELETE FROM verification_evidence/);
+    assert.match(db.execCalls[0], /GROUP BY task_id, slice_id, milestone_id, command, verdict/);
+  });
+
+  test("ensureVerificationEvidenceDedupIndex dedupes before creating the unique index", () => {
+    const db = new FakeAdapter();
+
+    ensureVerificationEvidenceDedupIndex(db);
+
+    assert.equal(db.execCalls.length, 2);
+    assert.match(db.execCalls[0], /DELETE FROM verification_evidence/);
+    assert.match(db.execCalls[1], /CREATE UNIQUE INDEX IF NOT EXISTS idx_verification_evidence_dedup/);
+  });
+
+  test("ensureVerificationEvidenceDedupIndex no-ops when the index already exists", () => {
+    const db = new FakeAdapter();
+    db.hasDedupIndex = true;
+
+    ensureVerificationEvidenceDedupIndex(db);
+
+    assert.deepEqual(db.execCalls, []);
+  });
+});

package/src/resources/extensions/gsd/tests/deep-project-auto-loop.test.ts

@@ -20,6 +20,7 @@ import {
   clearPendingAutoStart,
   checkDeepProjectSetupAfterTurn,
   clearPendingDeepProjectSetup,
+  FOREGROUND_DEEP_SETUP_RULE_NAMES,
   showSmartEntry,
   startDeepProjectSetupForeground,
 } from "../guided-flow.ts";
@@ -272,7 +273,16 @@ test("deep project setup: bootstrap can start auto-mode without an active milest
         shouldUseWorktreeIsolation: () => false,
         registerSigtermHandler: () => {},
         lockBase: () => base,
-        buildResolver: () => ({}) as any,
+        buildLifecycle: () => ({
+          adoptSessionRoot: (sessionBase: string, originalBase?: string) => {
+            s.basePath = sessionBase;
+            if (originalBase !== undefined) {
+              s.originalBasePath = originalBase;
+            } else if (!s.originalBasePath) {
+              s.originalBasePath = sessionBase;
+            }
+          },
+        }) as any,
       },
       {
         classification: "none",
@@ -377,7 +387,16 @@ test("deep project setup: bootstrap continues queued M002 without milestone cont
         shouldUseWorktreeIsolation: () => false,
         registerSigtermHandler: () => {},
         lockBase: () => base,
-        buildResolver: () => ({}) as any,
+        buildLifecycle: () => ({
+          adoptSessionRoot: (sessionBase: string, originalBase?: string) => {
+            s.basePath = sessionBase;
+            if (originalBase !== undefined) {
+              s.originalBasePath = originalBase;
+            } else if (!s.originalBasePath) {
+              s.originalBasePath = sessionBase;
+            }
+          },
+        }) as any,
       },
       {
         classification: "none",
@@ -1013,15 +1032,9 @@ test("deep project setup: same project advances when agent_end session id change
 });
 
 test("deep project setup: foreground dispatcher does not probe research-project rule", () => {
-  const source = readFileSync(new URL("../guided-flow.ts", import.meta.url), "utf-8");
-  const match = source.match(/const FOREGROUND_DEEP_SETUP_RULE_NAMES = new Set\(\[([\s\S]*?)\]\);/);
-  assert.ok(match, "foreground deep setup rule allowlist should be present");
-  assert.doesNotMatch(
-    match![1]!,
-    /research-project/,
-    "foreground setup must not evaluate research-project because that rule claims the inflight marker",
-  );
-  assert.match(source, /research-project have dispatch-time side effects/);
+  assert.equal(FOREGROUND_DEEP_SETUP_RULE_NAMES.has("deep: pre-planning (no PROJECT) → discuss-project"), true);
+  assert.equal(FOREGROUND_DEEP_SETUP_RULE_NAMES.has("deep: pre-planning (no research decision) → research-decision"), true);
+  assert.equal(FOREGROUND_DEEP_SETUP_RULE_NAMES.has("deep: pre-planning (no PROJECT research) → research-project"), false);
 });
 
 test("deep project setup: project-level units verify their real artifacts", () => {

package/src/resources/extensions/gsd/tests/deferred-milestone-dir-4996.test.ts

@@ -1,14 +1,10 @@
 // GSD Extension — Regression test for #4996: deferred milestone dir creation
-// Verifies that showHeadlessMilestoneCreation does not pre-create the milestone
-// directory before the discuss flow runs. The dir should only appear after a
-// writer (saveArtifactToDb / atomicWriteAsync) emits the first artifact.
 
 import { describe, it, beforeEach, afterEach } from "node:test";
 import assert from "node:assert/strict";
-import { mkdtempSync, mkdirSync, existsSync, rmSync, readFileSync } from "node:fs";
-import { dirname, join } from "node:path";
+import { mkdtempSync, mkdirSync, existsSync, rmSync } from "node:fs";
+import { join } from "node:path";
 import { tmpdir } from "node:os";
-import { fileURLToPath } from "node:url";
 
 import { isReusableGhostMilestone } from "../state.ts";
 import { nextMilestoneIdReserved } from "../milestone-id-reservation.ts";
@@ -16,47 +12,12 @@ import { clearReservedMilestoneIds, findMilestoneIds } from "../milestone-ids.ts
 import { invalidateAllCaches } from "../cache.ts";
 import { closeDatabase, openDatabase } from "../gsd-db.ts";
 
-const __dirname = dirname(fileURLToPath(import.meta.url));
-const GUIDED_FLOW_PATH = join(__dirname, "..", "guided-flow.ts");
-
-function getShowHeadlessBody(): string {
-  const source = readFileSync(GUIDED_FLOW_PATH, "utf-8");
-  const fnStart = source.indexOf("export async function showHeadlessMilestoneCreation");
-  assert.ok(fnStart > -1, "showHeadlessMilestoneCreation must exist in guided-flow.ts");
-  const nextExport = source.indexOf("\nexport ", fnStart + 1);
-  return source.slice(fnStart, nextExport === -1 ? source.length : nextExport);
-}
-
 function makeBase(prefix = "gsd-deferred-dir-"): string {
   const base = mkdtempSync(join(tmpdir(), prefix));
   mkdirSync(join(base, ".gsd", "milestones"), { recursive: true });
   return base;
 }
 
-describe("showHeadlessMilestoneCreation source guard (#4996)", () => {
-  it("does not call mkdirSync in the headless milestone creation path", () => {
-    const body = getShowHeadlessBody();
-    assert.doesNotMatch(
-      body,
-      /\bmkdirSync\s*\(/,
-      "showHeadlessMilestoneCreation must not pre-create milestone directories",
-    );
-  });
-
-  it("does not call mkdir or mkdirp before dispatchWorkflow", () => {
-    const body = getShowHeadlessBody();
-    const dispatchIdx = body.indexOf("dispatchWorkflow");
-    assert.ok(dispatchIdx > -1, "dispatchWorkflow must be present");
-
-    const beforeDispatch = body.slice(0, dispatchIdx);
-    assert.doesNotMatch(
-      beforeDispatch,
-      /\b(?:mkdir|mkdirp)\s*\(/,
-      "showHeadlessMilestoneCreation must defer directory creation until artifact write",
-    );
-  });
-});
-
 describe("deferred milestone dir creation (#4996)", () => {
   let base: string;
 
@@ -74,43 +35,31 @@ describe("deferred milestone dir creation (#4996)", () => {
   it("(a) fresh project: milestones dir has no M001 entry before any discuss flow", () => {
     base = makeBase();
     const nextId = nextMilestoneIdReserved(findMilestoneIds(base), false, base);
-    assert.equal(nextId, "M001", "reservation should choose M001 for a fresh project");
+    assert.equal(nextId, "M001");
 
     const ids = findMilestoneIds(base);
-    assert.equal(ids.length, 0, "no milestone dirs should exist before any discuss flow");
-
-    // And specifically M001 should not exist
-    const m001Dir = join(base, ".gsd", "milestones", "M001");
-    assert.ok(!existsSync(m001Dir), "M001 dir must not exist before the discuss flow runs");
+    assert.equal(ids.length, 0);
+    assert.ok(!existsSync(join(base, ".gsd", "milestones", "M001")));
   });
 
-  it("(b) abandoned discuss flow leaves no orphan: isReusableGhostMilestone returns false for non-existent dir", () => {
+  it("(b) abandoned discuss flow leaves no orphan", () => {
     base = makeBase();
     const nextId = nextMilestoneIdReserved(findMilestoneIds(base), false, base);
-    assert.equal(nextId, "M001", "reservation should not require a pre-created directory");
+    assert.equal(nextId, "M001");
 
     const m001Dir = join(base, ".gsd", "milestones", "M001");
-    assert.ok(!existsSync(m001Dir), "no M001 dir should exist");
-    assert.equal(isReusableGhostMilestone(base, "M001"), false, "non-existent milestone should not be reusable");
-    // findMilestoneIds only returns dirs that exist
-    const ids = findMilestoneIds(base);
-    assert.ok(!ids.includes("M001"), "M001 should not appear in findMilestoneIds when no dir exists");
+    assert.ok(!existsSync(m001Dir));
+    assert.equal(isReusableGhostMilestone(base, "M001"), false);
+    assert.ok(!findMilestoneIds(base).includes("M001"));
   });
 
-  it("(c) a stub dir left from a previous bug IS reusable but a newly-generated ID with no dir is not in the ghost list", () => {
+  it("(c) a stub dir left from a previous bug is reusable", () => {
     base = makeBase();
     openDatabase(join(base, ".gsd", "gsd.db"));
-    // Create a stub to represent a pre-existing phantom
     mkdirSync(join(base, ".gsd", "milestones", "M001", "slices"), { recursive: true });
 
-    // isReusableGhostMilestone identifies the orphaned stub
-    assert.ok(isReusableGhostMilestone(base, "M001"), "pre-existing stub should be identified as reusable ghost");
-    const nextId = nextMilestoneIdReserved(findMilestoneIds(base), false, base);
-    assert.equal(nextId, "M001", "reservation should reuse the pre-existing ghost");
-
-    // The new ID (M002, which would be max+1 in this scenario but ghost reuse returns M001)
-    // should not have a dir
-    const m002Dir = join(base, ".gsd", "milestones", "M002");
-    assert.ok(!existsSync(m002Dir), "a freshly-requested ID should have no dir until first artifact write");
+    assert.equal(isReusableGhostMilestone(base, "M001"), true);
+    assert.equal(nextMilestoneIdReserved(findMilestoneIds(base), false, base), "M001");
+    assert.ok(!existsSync(join(base, ".gsd", "milestones", "M002")));
   });
 });

package/src/resources/extensions/gsd/tests/detection.test.ts

@@ -11,12 +11,14 @@
 import test from "node:test";
 import assert from "node:assert/strict";
 import { mkdirSync, writeFileSync, rmSync, existsSync } from "node:fs";
+import { execFileSync } from "node:child_process";
 import { join } from "node:path";
 import { tmpdir } from "node:os";
 import {
   detectProjectState,
   detectV1Planning,
   detectProjectSignals,
+  classifyProject,
   scanProjectFiles,
 } from "../detection.ts";
 
@@ -37,6 +39,18 @@ function cleanup(dir: string): void {
   }
 }
 
+function git(dir: string, args: string[]): void {
+  execFileSync("git", args, { cwd: dir, stdio: "ignore" });
+}
+
+function makeGitRepo(prefix: string): string {
+  const dir = makeTempDir(prefix);
+  git(dir, ["init"]);
+  git(dir, ["config", "user.email", "test@example.com"]);
+  git(dir, ["config", "user.name", "Test User"]);
+  return dir;
+}
+
 // ─── detectProjectState ─────────────────────────────────────────────────────────
 
 test("detectProjectState: empty directory returns state=none", (t) => {
@@ -49,6 +63,132 @@ test("detectProjectState: empty directory returns state=none", (t) => {
   assert.equal(result.v2, undefined);
 });
 
+test("classifyProject: no git repo is invalid", (t) => {
+  const dir = makeTempDir("classify-invalid");
+  t.after(() => cleanup(dir));
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "invalid-repo");
+});
+
+test("classifyProject: empty git repo is greenfield", (t) => {
+  const dir = makeGitRepo("classify-greenfield");
+  t.after(() => cleanup(dir));
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "greenfield");
+});
+
+test("classifyProject: nested empty git repo does not inherit ancestor markers", (t) => {
+  const parent = makeGitRepo("classify-parent-marker");
+  t.after(() => cleanup(parent));
+
+  writeFileSync(join(parent, "package.json"), JSON.stringify({ name: "parent" }), "utf-8");
+  git(parent, ["add", "package.json"]);
+  git(parent, ["commit", "-m", "add parent marker"]);
+  const child = join(parent, "nested");
+  mkdirSync(child, { recursive: true });
+  git(child, ["init"]);
+  git(child, ["config", "user.email", "test@example.com"]);
+  git(child, ["config", "user.name", "Test User"]);
+
+  const classification = classifyProject(child);
+  assert.equal(classification.kind, "greenfield");
+});
+
+test("classifyProject: tracked static HTML is existing untyped content", (t) => {
+  const dir = makeGitRepo("classify-index");
+  t.after(() => cleanup(dir));
+
+  writeFileSync(join(dir, "index.html"), "<main></main>\n", "utf-8");
+  git(dir, ["add", "index.html"]);
+  git(dir, ["commit", "-m", "add static page"]);
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "untyped-existing");
+  assert.deepEqual(classification.contentFiles, ["index.html"]);
+});
+
+test("classifyProject: README-only repo is existing untyped content", (t) => {
+  const dir = makeGitRepo("classify-readme");
+  t.after(() => cleanup(dir));
+
+  writeFileSync(join(dir, "README.md"), "# docs\n", "utf-8");
+  git(dir, ["add", "README.md"]);
+  git(dir, ["commit", "-m", "add docs"]);
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "untyped-existing");
+});
+
+test("classifyProject: src-only content is untyped existing, not typed marker", (t) => {
+  const dir = makeGitRepo("classify-src-only");
+  t.after(() => cleanup(dir));
+
+  mkdirSync(join(dir, "src"), { recursive: true });
+  writeFileSync(join(dir, "src", "index.txt"), "content\n", "utf-8");
+  git(dir, ["add", "src/index.txt"]);
+  git(dir, ["commit", "-m", "add source content"]);
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "untyped-existing");
+  assert.deepEqual(classification.contentFiles, ["src/index.txt"]);
+});
+
+test("classifyProject: nested untracked files count as project content", (t) => {
+  const dir = makeGitRepo("classify-untracked-nested");
+  t.after(() => cleanup(dir));
+
+  mkdirSync(join(dir, "docs"), { recursive: true });
+  writeFileSync(join(dir, "docs", "index.html"), "<main></main>\n", "utf-8");
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "untyped-existing");
+  assert.deepEqual(classification.untrackedFiles, ["docs/index.html"]);
+});
+
+test("classifyProject: known markers produce typed existing project", (t) => {
+  const dir = makeGitRepo("classify-typed");
+  t.after(() => cleanup(dir));
+
+  writeFileSync(join(dir, "package.json"), JSON.stringify({ name: "typed" }), "utf-8");
+  git(dir, ["add", "package.json"]);
+  git(dir, ["commit", "-m", "add package"]);
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "typed-existing");
+  assert.ok(classification.markers.includes("package.json"));
+});
+
+test("classifyProject: ignored build/cache-only files do not count as content", (t) => {
+  const dir = makeGitRepo("classify-ignored");
+  t.after(() => cleanup(dir));
+
+  writeFileSync(join(dir, ".gitignore"), "dist/\n.cache/\n", "utf-8");
+  git(dir, ["add", ".gitignore"]);
+  git(dir, ["commit", "-m", "ignore generated files"]);
+  mkdirSync(join(dir, "dist"), { recursive: true });
+  writeFileSync(join(dir, "dist", "bundle.js"), "generated\n", "utf-8");
+  mkdirSync(join(dir, ".cache"), { recursive: true });
+  writeFileSync(join(dir, ".cache", "x"), "cache\n", "utf-8");
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "greenfield");
+});
+
+test("classifyProject: generated framework/cache dirs do not count as content", (t) => {
+  const dir = makeGitRepo("classify-generated-dirs");
+  t.after(() => cleanup(dir));
+
+  mkdirSync(join(dir, ".next", "server"), { recursive: true });
+  writeFileSync(join(dir, ".next", "server", "page.js"), "generated\n", "utf-8");
+  mkdirSync(join(dir, ".venv", "lib"), { recursive: true });
+  writeFileSync(join(dir, ".venv", "lib", "site.py"), "generated\n", "utf-8");
+
+  const classification = classifyProject(dir);
+  assert.equal(classification.kind, "greenfield");
+});
+
 test("detectProjectState: directory with .gsd/milestones/M001 returns v2-gsd", (t) => {
   const dir = makeTempDir("v2-gsd");
   t.after(() => cleanup(dir));

package/src/resources/extensions/gsd/tests/discuss-headless-rendering.test.ts

@@ -0,0 +1,37 @@
+// Project/App: GSD-2
+// File Purpose: Verifies the headless discussion prompt renders compact required guidance.
+
+import test from "node:test";
+import assert from "node:assert/strict";
+import { mkdtempSync, rmSync } from "node:fs";
+import { tmpdir } from "node:os";
+import { join } from "node:path";
+
+test("discuss-headless prompt renders compact investigation and audit guidance", async (t) => {
+  const previousGsdHome = process.env.GSD_HOME;
+  const providedGsdHome = process.env.GSD_TEST_HOME;
+  const isolatedHome = providedGsdHome ?? mkdtempSync(join(tmpdir(), "gsd-discuss-headless-render-"));
+  process.env.GSD_HOME = isolatedHome;
+  t.after(() => {
+    if (previousGsdHome === undefined) delete process.env.GSD_HOME;
+    else process.env.GSD_HOME = previousGsdHome;
+    if (!providedGsdHome) rmSync(isolatedHome, { recursive: true, force: true });
+  });
+
+  const { loadPrompt } = await import(`../prompt-loader.ts?test=${Date.now()}`);
+  const prompt = loadPrompt("discuss-headless", {
+    seedContext: "# Spec\n\nBuild the thing.",
+    milestoneId: "M001",
+    contextPath: ".gsd/milestones/M001/M001-CONTEXT.md",
+    commitInstruction: "Commit the created milestone artifacts.",
+    multiMilestoneCommitInstruction: "Commit the created milestone artifacts.",
+    inlinedTemplates: "## Template\n\nUse standard GSD artifacts.",
+  });
+
+  assert.match(prompt, /Investigate before making decisions:/);
+  assert.match(prompt, /Budget searches across investigation and focused research\./);
+  assert.match(prompt, /Resolve all of these from the spec and investigation before writing artifacts:/);
+  assert.match(prompt, /Print a structured depth summary in chat/);
+  assert.match(prompt, /Document every assumption in CONTEXT\.md/);
+  assert.doesNotMatch(prompt, /\{\{[a-zA-Z][a-zA-Z0-9_]*\}\}/);
+});

package/src/resources/extensions/gsd/tests/discuss-tool-scoping.test.ts

@@ -16,15 +16,12 @@
 
 import { describe, test } from "node:test";
 import assert from "node:assert/strict";
-import { readFileSync } from "node:fs";
-import { join, dirname } from "node:path";
-import { fileURLToPath } from "node:url";
+import { mkdtempSync, rmSync, writeFileSync } from "node:fs";
+import { join } from "node:path";
+import { tmpdir } from "node:os";
 
 import { DISCUSS_TOOLS_ALLOWLIST } from "../constants.ts";
-
-const __dirname = dirname(fileURLToPath(import.meta.url));
-const promptsDir = join(__dirname, "..", "prompts");
-const guidedFlowPath = join(__dirname, "..", "guided-flow.ts");
+import { _dispatchWorkflowForTest } from "../guided-flow.ts";
 
 // ─── Heavy tools that should NOT be in discuss scope ─────────────────────────
 
@@ -95,36 +92,46 @@ describe("discuss tool scoping (#2949)", () => {
     );
   });
 
-  test("guided-discuss-slice.md references gsd_summary_save", () => {
-    const prompt = readFileSync(join(promptsDir, "guided-discuss-slice.md"), "utf-8");
-    assert.ok(
-      prompt.includes("gsd_summary_save"),
-      "guided-discuss-slice.md should reference gsd_summary_save",
-    );
-  });
-
-  test("discuss.md references gsd_plan_milestone and gsd_decision_save", () => {
-    const prompt = readFileSync(join(promptsDir, "discuss.md"), "utf-8");
-    assert.ok(
-      prompt.includes("gsd_plan_milestone"),
-      "discuss.md should reference gsd_plan_milestone",
-    );
-    assert.ok(
-      prompt.includes("gsd_decision_save"),
-      "discuss.md should reference gsd_decision_save",
-    );
-  });
+  test("dispatchWorkflow scopes and restores tools for discuss unit types", async () => {
+    const originalWorkflowPath = process.env.GSD_WORKFLOW_PATH;
+    const tmp = mkdtempSync(join(tmpdir(), "gsd-discuss-tools-"));
+    const workflowPath = join(tmp, "GSD-WORKFLOW.md");
+    writeFileSync(workflowPath, "# Workflow\n");
+    const setCalls: string[][] = [];
+    const sent: unknown[] = [];
+    let sentTools: string[] = [];
+    let activeTools = ["gsd_summary_save", "gsd_complete_task", "bash"];
+    process.env.GSD_WORKFLOW_PATH = workflowPath;
+    try {
+      await _dispatchWorkflowForTest(
+        {
+          getActiveTools: () => [...activeTools],
+          setActiveTools: (tools: string[]) => {
+            setCalls.push([...tools]);
+            activeTools = [...tools];
+          },
+          sendMessage: (message: unknown) => {
+            sent.push(message);
+            sentTools = [...activeTools];
+          },
+        } as any,
+        "Discuss the project",
+        "gsd-run",
+        undefined,
+        "discuss-milestone",
+      );
+    } finally {
+      if (originalWorkflowPath === undefined) delete process.env.GSD_WORKFLOW_PATH;
+      else process.env.GSD_WORKFLOW_PATH = originalWorkflowPath;
+      rmSync(tmp, { recursive: true, force: true });
+    }
 
-  test("dispatchWorkflow source code scopes tools for discuss unit types", () => {
-    const source = readFileSync(guidedFlowPath, "utf-8");
-    // Verify that dispatchWorkflow references the allowlist for tool scoping
-    assert.ok(
-      source.includes("DISCUSS_TOOLS_ALLOWLIST"),
-      "guided-flow.ts should reference DISCUSS_TOOLS_ALLOWLIST for tool scoping",
-    );
-    assert.ok(
-      source.includes("setActiveTools"),
-      "guided-flow.ts should call setActiveTools to scope tools during discuss",
-    );
+    assert.deepEqual(setCalls[0], ["gsd_summary_save", "bash"]);
+    assert.deepEqual(setCalls.at(-1), ["gsd_summary_save", "gsd_complete_task", "bash"]);
+    assert.ok(sentTools.length > 0, "dispatch should queue a message");
+    assert.ok(sentTools.includes("gsd_summary_save"), "dispatch keeps the discuss save tool");
+    assert.ok(!sentTools.includes("gsd_complete_task"), "dispatch removes heavy completion tools");
+    assert.equal(sent.length, 1);
+    assert.match(String((sent[0] as { content?: unknown }).content), /Discuss the project/);
   });
 });