repro: windows hook flakiness

.github/workflows/chained_e2e.yml

@@ -275,7 +275,7 @@ jobs:
           UV_THREADPOOL_SIZE: '32'
           NODE_ENV: 'test'
         shell: 'pwsh'
-        run: 'npm run test:integration:sandbox:none'
+        run: 'npm run test:e2e'
 
   evals:
     name: 'Evals (ALWAYS_PASSING)'
@@ -315,19 +315,30 @@ jobs:
     needs:
       - 'e2e_linux'
       - 'e2e_mac'
+      - 'e2e_windows'
       - 'evals'
       - 'merge_queue_skipper'
     runs-on: 'gemini-cli-ubuntu-16-core'
     steps:
       - name: 'Check E2E test results'
         run: |
-          if [[ ${{ needs.e2e_linux.result }} != 'success' || \
-               ${{ needs.e2e_mac.result }} != 'success' || \
-               ${{ needs.evals.result }} != 'success' ]]; then
-            echo "One or more E2E jobs failed."
+          if [[ ${{ needs.e2e_linux.result }} != 'success' ]]; then
+            echo "Linux E2E job failed."
+            exit 1
+          fi
+          if [[ ${{ needs.e2e_mac.result }} != 'success' ]]; then
+            echo "macOS E2E job failed."
+            exit 1
+          fi
+          if [[ ${{ needs.e2e_windows.result }} != 'success' ]]; then
+            echo "Windows E2E job failed."
+            exit 1
+          fi
+          if [[ ${{ needs.evals.result }} != 'success' ]]; then
+            echo "Evals job failed."
             exit 1
           fi
-          echo "All required E2E jobs passed!"
+          echo "All E2E jobs passed!"
 
   set_workflow_status:
     runs-on: 'gemini-cli-ubuntu-16-core'

integration-tests/hooks-agent-flow.test.ts

@@ -5,7 +5,7 @@
  */
 
 import { describe, it, expect, beforeEach, afterEach } from 'vitest';
-import { TestRig } from './test-helper.js';
+import { TestRig, normalizePath } from './test-helper.js';
 import { join } from 'node:path';
 import { writeFileSync } from 'node:fs';
 
@@ -113,10 +113,9 @@ describe('Hooks Agent Flow', () => {
       }
       `;
 
-      const scriptPath = join(rig.testDir!, 'after_agent_verify.cjs');
-      writeFileSync(scriptPath, hookScript);
+      const scriptPath = rig.createScript('after_agent_verify.cjs', hookScript);
 
-      await rig.setup('should receive prompt and response in AfterAgent hook', {
+      rig.setup('should receive prompt and response in AfterAgent hook', {
         settings: {
           hooksConfig: {
             enabled: true,
@@ -127,7 +126,7 @@ describe('Hooks Agent Flow', () => {
                 hooks: [
                   {
                     type: 'command',
-                    command: `node "${scriptPath}"`,
+                    command: normalizePath(`node "${scriptPath}"`)!,
                     timeout: 5000,
                   },
                 ],
@@ -157,7 +156,7 @@ describe('Hooks Agent Flow', () => {
     });
 
     it('should process clearContext in AfterAgent hook output', async () => {
-      await rig.setup('should process clearContext in AfterAgent hook output', {
+      rig.setup('should process clearContext in AfterAgent hook output', {
         fakeResponsesPath: join(
           import.meta.dirname,
           'hooks-system.after-agent.responses',
@@ -171,18 +170,32 @@ describe('Hooks Agent Flow', () => {
         const input = JSON.parse(fs.readFileSync(0, 'utf-8'));
         const messageCount = input.llm_request?.contents?.length || 0;
         let counts = [];
-        try { counts = JSON.parse(fs.readFileSync('${messageCountFile}', 'utf-8')); } catch (e) {}
+        try { counts = JSON.parse(fs.readFileSync(${JSON.stringify(messageCountFile)}, 'utf-8')); } catch (e) {}
         counts.push(messageCount);
-        fs.writeFileSync('${messageCountFile}', JSON.stringify(counts));
+        fs.writeFileSync(${JSON.stringify(messageCountFile)}, JSON.stringify(counts));
         console.log(JSON.stringify({ decision: 'allow' }));
       `;
-      const beforeModelScriptPath = join(
-        rig.testDir!,
+      const beforeModelScriptPath = rig.createScript(
         'before_model_counter.cjs',
+        beforeModelScript,
       );
-      writeFileSync(beforeModelScriptPath, beforeModelScript);
 
-      await rig.setup('should process clearContext in AfterAgent hook output', {
+      const afterAgentScript = `
+        console.log(JSON.stringify({
+          decision: 'block',
+          reason: 'Security policy triggered',
+          hookSpecificOutput: {
+            hookEventName: 'AfterAgent',
+            clearContext: true
+          }
+        }));
+      `;
+      const afterAgentScriptPath = rig.createScript(
+        'after_agent_clear.cjs',
+        afterAgentScript,
+      );
+
+      rig.setup('should process clearContext in AfterAgent hook output', {
         settings: {
           hooks: {
             enabled: true,
@@ -191,7 +204,7 @@ describe('Hooks Agent Flow', () => {
                 hooks: [
                   {
                     type: 'command',
-                    command: `node "${beforeModelScriptPath}"`,
+                    command: normalizePath(`node "${beforeModelScriptPath}"`)!,
                     timeout: 5000,
                   },
                 ],
@@ -202,7 +215,7 @@ describe('Hooks Agent Flow', () => {
                 hooks: [
                   {
                     type: 'command',
-                    command: `node -e "console.log(JSON.stringify({decision: 'block', reason: 'Security policy triggered', hookSpecificOutput: {hookEventName: 'AfterAgent', clearContext: true}}))"`,
+                    command: normalizePath(`node "${afterAgentScriptPath}"`)!,
                     timeout: 5000,
                   },
                 ],
@@ -244,6 +257,22 @@ describe('Hooks Agent Flow', () => {
             import.meta.dirname,
             'hooks-agent-flow-multistep.responses',
           ),
+        },
+      );
+
+      // Create script files for hooks
+      const baPath = rig.createScript(
+        'ba_fired.cjs',
+        "console.log('BeforeAgent Fired');",
+      );
+      const aaPath = rig.createScript(
+        'aa_fired.cjs',
+        "console.log('AfterAgent Fired');",
+      );
+
+      await rig.setup(
+        'should fire BeforeAgent and AfterAgent exactly once per turn despite tool calls',
+        {
           settings: {
             hooksConfig: {
               enabled: true,
@@ -254,7 +283,7 @@ describe('Hooks Agent Flow', () => {
                   hooks: [
                     {
                       type: 'command',
-                      command: `node -e "console.log('BeforeAgent Fired')"`,
+                      command: normalizePath(`node "${baPath}"`)!,
                       timeout: 5000,
                     },
                   ],
@@ -265,7 +294,7 @@ describe('Hooks Agent Flow', () => {
                   hooks: [
                     {
                       type: 'command',
-                      command: `node -e "console.log('AfterAgent Fired')"`,
+                      command: normalizePath(`node "${aaPath}"`)!,
                       timeout: 5000,
                     },
                   ],