@opengsd/gsd-pi 1.2.0-dev.844675c9 → 1.2.0-dev.955e4da0

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

@@ -5,6 +5,10 @@ import { test } from 'node:test';
 import assert from 'node:assert/strict';
 import { registerDbTools } from '../bootstrap/db-tools.ts';
 import { getLegacyTelemetry, resetLegacyTelemetry } from '../legacy-telemetry.ts';
+import {
+  WORKFLOW_TOOL_ALIAS_PAIRS,
+  WORKFLOW_TOOL_CONTRACTS,
+} from '../workflow-tool-surface.ts';
 
 
 // ─── Mock PI ──────────────────────────────────────────────────────────────────
@@ -17,49 +21,38 @@ function makeMockPi() {
   } as any;
 }
 
-// ─── Rename map ───────────────────────────────────────────────────────────────
-
-const RENAME_MAP: Array<{ canonical: string; alias: string }> = [
-  { canonical: "gsd_decision_save", alias: "gsd_save_decision" },
-  { canonical: "gsd_requirement_update", alias: "gsd_update_requirement" },
-  { canonical: "gsd_requirement_save", alias: "gsd_save_requirement" },
-  { canonical: "gsd_summary_save", alias: "gsd_save_summary" },
-  { canonical: "gsd_milestone_generate_id", alias: "gsd_generate_milestone_id" },
-  { canonical: "gsd_task_complete", alias: "gsd_complete_task" },
-  { canonical: "gsd_slice_complete", alias: "gsd_complete_slice" },
-  { canonical: "gsd_plan_milestone", alias: "gsd_milestone_plan" },
-  { canonical: "gsd_plan_slice", alias: "gsd_slice_plan" },
-  { canonical: "gsd_plan_task", alias: "gsd_task_plan" },
-  { canonical: "gsd_replan_slice", alias: "gsd_slice_replan" },
-  { canonical: "gsd_reassess_roadmap", alias: "gsd_roadmap_reassess" },
-  { canonical: "gsd_complete_milestone", alias: "gsd_milestone_complete" },
-  { canonical: "gsd_validate_milestone", alias: "gsd_milestone_validate" },
-  { canonical: "gsd_task_reopen", alias: "gsd_reopen_task" },
-  { canonical: "gsd_slice_reopen", alias: "gsd_reopen_slice" },
-  { canonical: "gsd_milestone_reopen", alias: "gsd_reopen_milestone" },
-];
-
-const STANDALONE_TOOLS = [
-  "gsd_save_gate_result",
-  "gsd_skip_slice",
-  "gsd_uat_result_save",
-];
-
 // ─── Registration count ──────────────────────────────────────────────────────
 
 console.log('\n── Tool naming: registration count ──');
 
 const pi = makeMockPi();
 registerDbTools(pi);
-
+const toolByName = new Map<string, any>(pi.tools.map((tool: any) => [tool.name, tool]));
+const registeredCanonicalNames = new Set<string>(
+  WORKFLOW_TOOL_CONTRACTS
+    .map((tool) => tool.canonicalName)
+    .filter((name) => toolByName.has(name)),
+);
+const RENAME_MAP = WORKFLOW_TOOL_ALIAS_PAIRS.filter(({ canonical }) =>
+  registeredCanonicalNames.has(canonical),
+);
+const STANDALONE_TOOLS = WORKFLOW_TOOL_CONTRACTS
+  .filter((tool) => registeredCanonicalNames.has(tool.canonicalName) && tool.aliases.length === 0)
+  .map((tool) => tool.canonicalName);
+const expectedRegisteredNames = [
+  ...STANDALONE_TOOLS,
+  ...RENAME_MAP.flatMap(({ canonical, alias }) => [canonical, alias]),
+].sort();
+
+assert.equal(pi.tools.length, toolByName.size, 'Tool registration should not produce duplicate names');
 assert.deepStrictEqual(
-  pi.tools.length,
-  RENAME_MAP.length * 2 + STANDALONE_TOOLS.length,
-  'Should register canonical/alias tool pairs plus standalone DB tools',
+  [...toolByName.keys()].sort(),
+  expectedRegisteredNames,
+  'Should register only workflow surface tools and their declared aliases',
 );
 
 for (const name of STANDALONE_TOOLS) {
-  assert.ok(pi.tools.some((t: any) => t.name === name), `Standalone tool "${name}" should be registered`);
+  assert.ok(toolByName.has(name), `Standalone tool "${name}" should be registered`);
 }
 
 // ─── Both names exist for each pair ──────────────────────────────────────────
@@ -67,8 +60,8 @@ for (const name of STANDALONE_TOOLS) {
 console.log('\n── Tool naming: canonical and alias names exist ──');
 
 for (const { canonical, alias } of RENAME_MAP) {
-  const canonicalTool = pi.tools.find((t: any) => t.name === canonical);
-  const aliasTool = pi.tools.find((t: any) => t.name === alias);
+  const canonicalTool = toolByName.get(canonical);
+  const aliasTool = toolByName.get(alias);
 
   assert.ok(canonicalTool !== undefined, `Canonical tool "${canonical}" should be registered`);
   assert.ok(aliasTool !== undefined, `Alias tool "${alias}" should be registered`);
@@ -79,8 +72,8 @@ for (const { canonical, alias } of RENAME_MAP) {
 console.log('\n── Tool naming: alias execute wrapper ──');
 
 for (const { canonical, alias } of RENAME_MAP) {
-  const canonicalTool = pi.tools.find((t: any) => t.name === canonical);
-  const aliasTool = pi.tools.find((t: any) => t.name === alias);
+  const canonicalTool = toolByName.get(canonical);
+  const aliasTool = toolByName.get(alias);
 
   if (canonicalTool && aliasTool) {
     assert.ok(
@@ -91,8 +84,8 @@ for (const { canonical, alias } of RENAME_MAP) {
 }
 
 test("alias execute increments legacy MCP alias telemetry before delegating", async () => {
-  const canonicalTool = pi.tools.find((t: any) => t.name === "gsd_decision_save");
-  const aliasTool = pi.tools.find((t: any) => t.name === "gsd_save_decision");
+  const canonicalTool = toolByName.get("gsd_decision_save");
+  const aliasTool = toolByName.get("gsd_save_decision");
   assert.ok(canonicalTool);
   assert.ok(aliasTool);
 
@@ -120,7 +113,7 @@ test("alias execute increments legacy MCP alias telemetry before delegating", as
 console.log('\n── Tool naming: alias descriptions ──');
 
 for (const { canonical, alias } of RENAME_MAP) {
-  const aliasTool = pi.tools.find((t: any) => t.name === alias);
+  const aliasTool = toolByName.get(alias);
 
   if (aliasTool) {
     assert.ok(
@@ -135,7 +128,7 @@ for (const { canonical, alias } of RENAME_MAP) {
 console.log('\n── Tool naming: canonical promptGuidelines use canonical name ──');
 
 for (const { canonical } of RENAME_MAP) {
-  const canonicalTool = pi.tools.find((t: any) => t.name === canonical);
+  const canonicalTool = toolByName.get(canonical);
 
   if (canonicalTool) {
     const guidelinesText = canonicalTool.promptGuidelines.join(' ');
@@ -151,7 +144,7 @@ for (const { canonical } of RENAME_MAP) {
 console.log('\n── Tool naming: alias promptGuidelines redirect to canonical ──');
 
 for (const { canonical, alias } of RENAME_MAP) {
-  const aliasTool = pi.tools.find((t: any) => t.name === alias);
+  const aliasTool = toolByName.get(alias);
 
   if (aliasTool) {
     const guidelinesText = aliasTool.promptGuidelines.join(' ');

package/src/resources/extensions/gsd/tests/uat-policy.test.ts

@@ -66,6 +66,9 @@ describe("uat-policy", () => {
   it("detects direct and MCP-shaped browser tool surfaces", () => {
     assert.equal(hasUatBrowserToolSurface(["read", "browser_navigate"]), true);
     assert.equal(hasUatBrowserToolSurface(["read", "mcp__gsd-browser__browser_navigate"]), true);
+    assert.equal(hasUatBrowserToolSurface(["read", "mcp__gsd-browser__*"]), true);
+    assert.equal(hasUatBrowserToolSurface(["read", "mcp__browser-uat__*"]), true);
+    assert.equal(hasUatBrowserToolSurface(["read", "mcp__gsd-workflow__*"]), false);
     assert.equal(hasUatBrowserToolSurface(["read", "gsd_uat_exec"]), false);
     assert.equal(hasUatBrowserToolSurface(undefined), false);
   });
@@ -89,6 +92,26 @@ describe("uat-policy", () => {
       }),
       null,
     );
+    assert.equal(
+      getUatBrowserToolSupportError({
+        uatType: "human-experience",
+        activeTools: ["read", "gsd_uat_exec"],
+        registeredTools: ["browser_navigate"],
+        milestoneId: "M001",
+        sliceId: "S01",
+      }),
+      null,
+    );
+    assert.equal(
+      getUatBrowserToolSupportError({
+        uatType: "human-experience",
+        activeTools: ["read", "gsd_uat_exec"],
+        registeredTools: ["mcp__gsd-browser__*"],
+        milestoneId: "M001",
+        sliceId: "S01",
+      }),
+      null,
+    );
 
     const error = getUatBrowserToolSupportError({
       uatType: "browser-executable",

package/src/resources/extensions/gsd/tests/unit-context-composer.test.ts

@@ -8,6 +8,7 @@ import { join } from "node:path";
 import { tmpdir } from "node:os";
 
 import {
+  composeContractedUnitContext,
   composeContextModeInstructions,
   composeInlinedContext,
   composeUnitContext,
@@ -15,6 +16,7 @@ import {
   type ArtifactResolver,
   type ExcerptResolver,
 } from "../unit-context-composer.ts";
+import { compileUnitContextContract } from "../tool-contract.ts";
 import type {
   ArtifactKey,
   BaseResolverContext,
@@ -268,6 +270,21 @@ test("#4782 phase 2: buildReassessRoadmapPrompt emits composer-shaped context wi
   assert.ok(!prompt.includes("Slice Context (from discussion)"));
 });
 
+test("execute-task prompt surfaces contract-declared on-demand slice research", async (t) => {
+  const base = makeFixtureBase();
+  t.after(() => cleanup(base));
+  invalidateAllCaches();
+
+  seed(base, "M001");
+  writeArtifacts(base);
+
+  const prompt = await buildExecuteTaskPrompt("M001", "S01", "First", "T01", "Task", base);
+
+  assert.match(prompt, /## On-demand Context/);
+  assert.match(prompt, /\.gsd\/milestones\/M001\/slices\/S01\/S01-RESEARCH\.md/);
+  assert.match(prompt, /Read it only if the inlined task plan, slice plan excerpt, and carry-forward context do not explain/);
+});
+
 test("Context Mode resume injection: eligible prompts include one bounded snapshot block above inlined context", async (t) => {
   const base = makeFixtureBase();
   t.after(() => cleanup(base));
@@ -370,6 +387,36 @@ test("#4924 v2 composer: returns empty sections for unknown unit type", async ()
   assert.deepEqual(out, { prepend: "", inline: "" });
 });
 
+test("Unit Context Contract composer exposes keyed blocks and on-demand artifacts", async () => {
+  const result = compileUnitContextContract("execute-task");
+  assert.equal(result.ok, true);
+  if (!result.ok) return;
+
+  const calls: ArtifactKey[] = [];
+  const out = await composeContractedUnitContext(result.contract, {
+    base: { ...fakeBase, unitType: "stale-unit", taskId: "T01" },
+    resolveArtifact: async (key) => {
+      calls.push(key);
+      return `BODY:${key}`;
+    },
+  });
+
+  assert.deepEqual(calls, [
+    "task-plan",
+    "slice-plan",
+    "prior-task-summaries",
+    "templates",
+  ]);
+  assert.deepEqual(out.blocks.map((block) => [block.key, block.mode]), [
+    ["task-plan", "inline"],
+    ["slice-plan", "inline"],
+    ["prior-task-summaries", "inline"],
+    ["templates", "inline"],
+  ]);
+  assert.deepEqual(out.onDemand, ["slice-research"]);
+  assert.match(out.inline, /BODY:task-plan\n\n---\n\nBODY:slice-plan/);
+});
+
 test("#4924 v2 composer: omitting resolveArtifact skips inline keys without erroring", async () => {
   const out = await composeUnitContext("reassess-roadmap", { base: fakeBase });
   assert.strictEqual(out.inline, "");

package/src/resources/extensions/gsd/tests/user-input-boundary.test.ts

@@ -2,7 +2,11 @@ import test from "node:test";
 import assert from "node:assert/strict";
 
 import * as userInputBoundary from "../user-input-boundary.ts";
-import { isAwaitingUserInput, shouldPauseForUserApprovalQuestion } from "../user-input-boundary.ts";
+import {
+  isAwaitingUserInput,
+  messageHasPendingAskUserQuestionsTool,
+  shouldPauseForUserApprovalQuestion,
+} from "../user-input-boundary.ts";
 
 test("lastAssistantText extracts the latest assistant text block content", () => {
   const lastAssistantText = (userInputBoundary as {
@@ -73,6 +77,87 @@ test("isAwaitingUserInput does not trigger on thinking-block approval phrases",
   assert.equal(shouldPauseForUserApprovalQuestion("discuss-requirements", messages), false);
 });
 
+test("messageHasPendingAskUserQuestionsTool detects in-flight structured question tools", () => {
+  // state: "running" with no externalResult → still in-flight
+  assert.equal(
+    messageHasPendingAskUserQuestionsTool({
+      role: "assistant",
+      content: [
+        { type: "text", text: "Which direction?" },
+        { type: "toolCall", name: "mcp__gsd-workflow__ask_user_questions", state: "running" },
+      ],
+    }),
+    true,
+  );
+
+  // no state, no externalResult — streaming block that hasn't completed yet
+  assert.equal(
+    messageHasPendingAskUserQuestionsTool({
+      role: "assistant",
+      content: [
+        { type: "toolCall", name: "ask_user_questions" },
+      ],
+    }),
+    true,
+  );
+
+  // state: "completed" — legacy state-based completion
+  assert.equal(
+    messageHasPendingAskUserQuestionsTool({
+      role: "assistant",
+      content: [
+        { type: "toolCall", name: "ask_user_questions", state: "completed" },
+      ],
+    }),
+    false,
+  );
+
+  // externalResult present — Claude Code signals completion via externalResult, not state
+  assert.equal(
+    messageHasPendingAskUserQuestionsTool({
+      role: "assistant",
+      content: [
+        { type: "toolCall", name: "ask_user_questions", externalResult: { content: [], isError: false } },
+      ],
+    }),
+    false,
+  );
+  assert.equal(
+    messageHasPendingAskUserQuestionsTool({
+      role: "assistant",
+      content: [
+        {
+          type: "toolCall",
+          name: "ask_user_questions",
+          externalResult: { content: [{ type: "text", text: "answer" }], isError: false },
+        },
+      ],
+    }),
+    false,
+  );
+
+  // serverToolUse shape (claude-code-cli MCP path) — no externalResult → in-flight
+  assert.equal(
+    messageHasPendingAskUserQuestionsTool({
+      role: "assistant",
+      content: [
+        { type: "serverToolUse", name: "mcp__gsd-workflow__ask_user_questions" },
+      ],
+    }),
+    true,
+  );
+  // serverToolUse shape — externalResult present → completed
+  assert.equal(
+    messageHasPendingAskUserQuestionsTool({
+      role: "assistant",
+      content: [
+        { type: "serverToolUse", name: "ask_user_questions", externalResult: { content: [], isError: false } },
+      ],
+    }),
+    false,
+  );
+});
+
 test("isAwaitingUserInput still triggers on text-block question marks when thinking is also present", () => {
   // When thinking + text are both present and the text asks a question, it should still pause.
   const messages = [

package/src/resources/extensions/gsd/tests/validate-milestone-stuck-guard.test.ts

@@ -12,6 +12,7 @@ import { clearPathCache } from "../paths.ts";
 import {
   openDatabase,
   closeDatabase,
+  insertAssessment,
   insertMilestone,
   insertSlice,
 } from "../gsd-db.ts";
@@ -222,6 +223,44 @@ describe("validate-milestone stuck-loop guard (#4094)", () => {
     assert.equal(s.pendingVerificationRetry, null);
   });
 
+  test("continues when DB pass overrides stale worktree needs-attention after /gsd verdict", async () => {
+    insertMilestone({ id: "M001" });
+    insertSlice({ id: "S01", milestoneId: "M001", title: "Slice 1", status: "complete" });
+    writeWorktreeValidationFile("needs-attention");
+
+    const worktreeValidationPath = join(
+      tempDir,
+      ".gsd",
+      "worktrees",
+      "M001",
+      ".gsd",
+      "milestones",
+      "M001",
+      "M001-VALIDATION.md",
+    );
+    insertAssessment({
+      path: worktreeValidationPath,
+      milestoneId: "M001",
+      sliceId: null,
+      taskId: null,
+      status: "pass",
+      scope: "milestone-validation",
+      fullContent: "---\nverdict: pass\n---\n\n# Validation\nManually overridden via /gsd verdict\n",
+    });
+    invalidateAllCaches();
+
+    const ctx = makeMockCtx();
+    const pi = makeMockPi();
+    const pauseAutoMock = mock.fn(async () => {});
+    const s = makeMockSession(tempDir, "validate-milestone", "M001");
+
+    const result = await runPostUnitVerification({ s, ctx, pi } as VerificationContext, pauseAutoMock);
+
+    assert.equal(result, "continue");
+    assert.equal(pauseAutoMock.mock.callCount(), 0);
+    assert.equal(ctx.ui.notify.mock.callCount(), 0);
+  });
+
   test("retries when no VALIDATION file exists yet", async () => {
     insertMilestone({ id: "M001" });
     insertSlice({ id: "S01", milestoneId: "M001", title: "Slice 1", status: "complete" });

package/src/resources/extensions/gsd/tests/web-app-uat.test.ts

@@ -0,0 +1,150 @@
+import { describe, test } from "node:test";
+import assert from "node:assert/strict";
+import { mkdirSync, mkdtempSync, rmSync, writeFileSync } from "node:fs";
+import { join } from "node:path";
+import { tmpdir } from "node:os";
+
+import {
+  buildWebAppUatGuidanceBlock,
+  detectWebApp,
+  findPlaywrightTestScript,
+  hasPlaywrightTestDependency,
+} from "../web-app-uat.ts";
+
+function scaffoldProject(root: string, pkg: Record<string, unknown>): void {
+  writeFileSync(join(root, "package.json"), JSON.stringify(pkg, null, 2));
+}
+
+describe("web-app-uat guidance", () => {
+  test("returns null for non-web projects", () => {
+    const root = mkdtempSync(join(tmpdir(), "gsd-web-uat-"));
+    try {
+      writeFileSync(join(root, "README.md"), "# CLI tool\n");
+      assert.equal(detectWebApp(root), false);
+      assert.equal(buildWebAppUatGuidanceBlock(root), null);
+    } finally {
+      rmSync(root, { recursive: true, force: true });
+    }
+  });
+
+  test("returns guidance for react/vite web apps", () => {
+    const root = mkdtempSync(join(tmpdir(), "gsd-web-uat-"));
+    try {
+      scaffoldProject(root, {
+        dependencies: { react: "19.0.0", "react-dom": "19.0.0" },
+        devDependencies: { vite: "6.0.0" },
+        scripts: { dev: "vite" },
+      });
+      assert.equal(detectWebApp(root), true);
+      const block = buildWebAppUatGuidanceBlock(root);
+      assert.ok(block);
+      assert.match(block!, /browser-executable/);
+      assert.match(block!, /Playwright scaffolding/);
+    } finally {
+      rmSync(root, { recursive: true, force: true });
+    }
+  });
+
+  test("detects existing Playwright and npm script", () => {
+    const root = mkdtempSync(join(tmpdir(), "gsd-web-uat-"));
+    try {
+      scaffoldProject(root, {
+        dependencies: { next: "15.0.0" },
+        devDependencies: { "@playwright/test": "1.60.0", playwright: "1.60.0" },
+        scripts: { "test:e2e": "playwright test" },
+      });
+      assert.equal(hasPlaywrightTestDependency(root), true);
+      assert.equal(findPlaywrightTestScript(root), "npm run test:e2e");
+      const block = buildWebAppUatGuidanceBlock(root);
+      assert.ok(block);
+      assert.match(block!, /dependency detected/);
+      assert.match(block!, /npm run test:e2e/);
+      assert.doesNotMatch(block!, /Playwright scaffolding/);
+    } finally {
+      rmSync(root, { recursive: true, force: true });
+    }
+  });
+
+  test("does not mistake a non-playwright e2e script as playwright", () => {
+    const root = mkdtempSync(join(tmpdir(), "gsd-web-uat-"));
+    try {
+      scaffoldProject(root, {
+        dependencies: { react: "19.0.0" },
+        devDependencies: { "@playwright/test": "1.60.0" },
+        scripts: { e2e: "cypress run", dev: "vite" },
+      });
+      // The e2e script runs Cypress, not Playwright — must not be returned
+      assert.equal(findPlaywrightTestScript(root), null);
+      const block = buildWebAppUatGuidanceBlock(root);
+      assert.ok(block);
+      // Playwright dep is present so guidance shows "dependency detected", not "scaffolding"
+      assert.match(block!, /dependency detected/);
+      assert.doesNotMatch(block!, /Playwright scaffolding/);
+      // Falls back to generic npx command because no playwright-named script exists
+      assert.doesNotMatch(block!, /npm run e2e/);
+    } finally {
+      rmSync(root, { recursive: true, force: true });
+    }
+  });
+
+  test("recognises playwright via script value even without a local dependency", () => {
+    const root = mkdtempSync(join(tmpdir(), "gsd-web-uat-"));
+    try {
+      scaffoldProject(root, {
+        dependencies: { next: "15.0.0" },
+        scripts: { "test:e2e": "npx playwright test" },
+      });
+      assert.equal(hasPlaywrightTestDependency(root), false);
+      assert.equal(findPlaywrightTestScript(root), "npm run test:e2e");
+      const block = buildWebAppUatGuidanceBlock(root);
+      assert.ok(block);
+      // Script-based detection should trigger "dependency detected" path, not scaffolding
+      assert.match(block!, /dependency detected/);
+      assert.match(block!, /npm run test:e2e/);
+      assert.doesNotMatch(block!, /Playwright scaffolding/);
+    } finally {
+      rmSync(root, { recursive: true, force: true });
+    }
+  });
+
+  test("does not treat playwright install script as a test command", () => {
+    const root = mkdtempSync(join(tmpdir(), "gsd-web-uat-"));
+    try {
+      scaffoldProject(root, {
+        dependencies: { react: "19.0.0" },
+        devDependencies: { "@playwright/test": "1.60.0" },
+        // postinstall runs 'playwright install' (browser setup), not a test
+        scripts: { postinstall: "playwright install", "test:e2e": "npx playwright test" },
+      });
+      // install script must not be returned; the real test script should be
+      assert.equal(findPlaywrightTestScript(root), "npm run test:e2e");
+    } finally {
+      rmSync(root, { recursive: true, force: true });
+    }
+  });
+
+  test("returns null when only an install script mentions playwright", () => {
+    const root = mkdtempSync(join(tmpdir(), "gsd-web-uat-"));
+    try {
+      scaffoldProject(root, {
+        dependencies: { react: "19.0.0" },
+        devDependencies: { "@playwright/test": "1.60.0" },
+        scripts: { postinstall: "playwright install --with-deps" },
+      });
+      assert.equal(findPlaywrightTestScript(root), null);
+    } finally {
+      rmSync(root, { recursive: true, force: true });
+    }
+  });
+
+  test("detects static sites via index.html", () => {
+    const root = mkdtempSync(join(tmpdir(), "gsd-web-uat-"));
+    try {
+      mkdirSync(join(root, "public"), { recursive: true });
+      writeFileSync(join(root, "public", "index.html"), "<html></html>");
+      assert.equal(detectWebApp(root), true);
+    } finally {
+      rmSync(root, { recursive: true, force: true });
+    }
+  });
+});