fix(runner): align SDK pin and complete real runtime verification (#2525)

* fix(runner): align SDK pin and workspace-aware integration fixtures

* fix(ci): format sources and resolve current migration head

* test(persistence): align standalone migration fixtures with current models

* test(web): align smoke fixtures with current processor UI

---------

Co-authored-by: dadachann <185672915+dadachann@users.noreply.github.com>

web/tests/e2e/crud-smoke.spec.ts

@@ -89,7 +89,7 @@ test.describe('frontend CRUD smoke flows', () => {
     await expect(
       page.locator('[data-processor-kind="pipeline"]'),
     ).toContainText(
-      '流水线即为经典的“收到消息、请求AI、回复用户”流程，并辅以常用的配置功能。仅处理消息事件，适合步骤明确、需要稳定控制处理过程的场景。',
+      '流水线按“接收消息、调用 AI、回复用户”的固定流程运行，可配置知识库和插件扩展。仅处理消息事件，适合步骤明确、需要控制处理过程的场景。',
     );
   });
 
@@ -1034,7 +1034,7 @@ test.describe('agent runner resource selectors', () => {
     await expect(
       page.getByText('No Runner extension is installed yet.'),
     ).toBeVisible();
-    await expect(page.getByText('Runner Marketplace')).toBeVisible();
+    await expect(page.getByText('Runner plugins in Marketplace')).toBeVisible();
     const selectorPopup = page.locator('[data-slot="select-content"]');
     await expect(
       selectorPopup.getByText('Runner used by the grouped selector test.', {
@@ -1055,20 +1055,35 @@ test.describe('agent runner resource selectors', () => {
       });
 
     await page
-      .getByRole('option')
-      .filter({ hasText: 'Marketplace Runner' })
+      .getByRole('button', { name: 'Install Marketplace Runner', exact: true })
       .click();
 
     await expect.poll(() => installRequests).toBe(1);
     await expect.poll(() => taskPolls).toBeGreaterThan(0);
+    await expect(
+      page.getByRole('option', {
+        name: 'Marketplace Runner Runner used by the grouped selector test.',
+        exact: true,
+      }),
+    ).toBeVisible();
+    await page
+      .getByRole('option', {
+        name: 'Marketplace Runner Runner used by the grouped selector test.',
+        exact: true,
+      })
+      .click();
     await expect(runnerSelect).toContainText('Marketplace Runner');
 
     await runnerSelect.click();
     await expect(
-      page.getByRole('option').filter({ hasText: runnerId }),
+      page.getByRole('option', {
+        name: 'Marketplace Runner Runner used by the grouped selector test.',
+        exact: true,
+      }),
     ).toBeVisible();
     await expect(
-      page.getByText('Runner used by the grouped selector test.', {
+      page.getByRole('button', {
+        name: 'Install Marketplace Runner',
         exact: true,
       }),
     ).toHaveCount(0);
@@ -1130,7 +1145,7 @@ test.describe('agent runner resource selectors', () => {
     const runnerSelect = page.getByRole('combobox', { name: 'Runner' });
     await runnerSelect.click();
     await expect(page.getByText('Installed Runners')).toBeVisible();
-    await expect(page.getByText('Runner Marketplace')).toBeVisible();
+    await expect(page.getByText('Runner plugins in Marketplace')).toBeVisible();
     await expect(
       page
         .locator('[data-slot="select-content"]')
@@ -1139,9 +1154,10 @@ test.describe('agent runner resource selectors', () => {
         }),
     ).toBeVisible();
     await expect(
-      page
-        .getByRole('option')
-        .filter({ hasText: 'Pipeline Marketplace Runner' }),
+      page.getByRole('button', {
+        name: 'Install Pipeline Marketplace Runner',
+        exact: true,
+      }),
     ).toBeVisible();
   });
 
@@ -1214,8 +1230,11 @@ test.describe('agent and pipeline save concurrency', () => {
 
     await page.goto('/home/agents?id=agent-save-race');
     const saveButton = page.getByRole('button', { name: /^Save$/ });
-    await page.getByRole('tab', { name: 'Bindable Event Range' }).click();
-    const eventPatterns = page.getByLabel('Event Range');
+    await page.getByRole('tab', { name: 'Events & tools' }).click();
+    const eventPatterns = page.getByRole('button', {
+      name: 'Add event',
+      exact: true,
+    });
     await expect(eventPatterns).toBeVisible();
 
     await eventPatterns.click();
@@ -1231,8 +1250,8 @@ test.describe('agent and pipeline save concurrency', () => {
     await eventPatterns.click();
     await page.getByRole('option').filter({ hasText: 'group.*' }).click();
     await page
-      .getByRole('option')
-      .filter({ hasText: 'message.received' })
+      .getByRole('button', { name: 'Remove event', exact: true })
+      .first()
       .click();
     await page.keyboard.press('Escape');
     await forceFormSubmit(page, '#agent-form');

web/tests/e2e/fixtures/langbot-api.ts

@@ -725,10 +725,12 @@ async function handleBackendApi(route: Route, state: LangBotApiMockState) {
     return fulfillJson(route, { agents: state.pipelines });
   }
 
-  const agentDebugMatch = path.match(/^\/api\/v1\/agents\/([^/]+)\/debug$/);
+  const agentDebugMatch = path.match(
+    /^\/api\/v1\/agents\/([^/]+)\/debug(?:\/stream)?$/,
+  );
   if (agentDebugMatch) {
     const payload = parseJsonBody(route);
-    return fulfillJson(route, {
+    const result = {
       event_id: nextId(state, 'event'),
       event_type: String(payload.event_type || 'message.received'),
       conversation_id: String(payload.conversation_id || 'debug-session'),
@@ -740,7 +742,15 @@ async function handleBackendApi(route: Route, state: LangBotApiMockState) {
           text: 'Mock Agent response',
         },
       ],
-    });
+    };
+    if (path.endsWith('/stream')) {
+      return route.fulfill({
+        status: 200,
+        contentType: 'application/x-ndjson',
+        body: JSON.stringify({ kind: 'completed', data: result }) + '\n',
+      });
+    }
+    return fulfillJson(route, result);
   }
 
   const agentMatch = path.match(/^\/api\/v1\/agents\/([^/]+)$/);

web/tests/e2e/home-smoke.spec.ts

@@ -11,7 +11,7 @@ const appRoutes = [
   {
     path: '/home/agents',
     heading: 'Processors',
-    bodyText: 'Select an Agent or Pipeline from the sidebar',
+    bodyText: 'Select a processor from the sidebar',
   },
   {
     path: '/home/extensions',

web/tests/e2e/processor-detail-workbench.spec.ts

@@ -60,8 +60,8 @@ test.describe('processor detail workbench', () => {
 
     const appShell = page.locator('[class*="group/sidebar-wrapper"]');
     const sidebarInset = page.locator('[data-slot="sidebar-inset"]');
-    await expect(appShell).toHaveCSS('overflow', 'clip');
-    await expect(sidebarInset).toHaveCSS('overflow', 'clip');
+    await expect(appShell).toHaveCSS('overflow', 'hidden');
+    await expect(sidebarInset).toHaveCSS('overflow', 'hidden');
     await appShell.evaluate((element) => {
       element.scrollTop = 300;
     });
@@ -79,9 +79,7 @@ test.describe('processor detail workbench', () => {
     const flow = configPanel.getByRole('tablist');
     await expect(flow.getByRole('tab').nth(0)).toContainText('Runner');
     await expect(flow.getByRole('tab').nth(1)).toContainText('Local Agent');
-    await expect(flow.getByRole('tab').nth(2)).toContainText(
-      'Bindable Event Range',
-    );
+    await expect(flow.getByRole('tab').nth(2)).toContainText('Events & tools');
     await expect(flow.getByRole('tab')).toHaveCount(3);
     await expect(flow.getByText('Management')).toHaveCount(0);
 
@@ -126,19 +124,32 @@ test.describe('processor detail workbench', () => {
 
     await flow.getByRole('tab').nth(2).click();
     await expect(
-      configPanel.getByText('Bindable Event Range', { exact: true }).last(),
+      configPanel.getByText('Events & tools', { exact: true }).last(),
     ).toBeVisible();
-    const eventPicker = configPanel.getByRole('combobox', {
-      name: 'Event Range',
+    const eventPicker = configPanel.getByRole('button', {
+      name: 'Add event',
+      exact: true,
     });
     await expect(eventPicker).toBeVisible();
-    await expect(configPanel.getByRole('textbox')).toHaveCount(0);
+    await expect(configPanel.getByRole('textbox')).toHaveCount(1);
+    await expect(
+      configPanel.getByRole('textbox', { name: 'Search tools…' }),
+    ).toBeVisible();
     await eventPicker.click();
     await expect(
       page.getByRole('option').filter({ hasText: 'message.received' }),
     ).toBeVisible();
-    await expect(page.getByRole('group', { name: 'Messages' })).toHaveCount(1);
-    await expect(page.getByRole('group', { name: 'Groups' })).toHaveCount(1);
+    await expect(
+      page.getByRole('option').filter({ hasText: 'group.*' }),
+    ).toBeVisible();
+    await expect(
+      page
+        .locator('[cmdk-group-heading]')
+        .getByText('Messages', { exact: true }),
+    ).toBeVisible();
+    await expect(
+      page.locator('[cmdk-group-heading]').getByText('Groups', { exact: true }),
+    ).toBeVisible();
     await page
       .getByRole('option')
       .filter({ hasText: 'message.*' })
@@ -189,7 +200,7 @@ test.describe('processor detail workbench', () => {
       }
       if (
         request.method() === 'POST' &&
-        path === '/api/v1/agents/agent-workbench/debug'
+        path === '/api/v1/agents/agent-workbench/debug/stream'
       ) {
         requests.push('debug');
       }
@@ -237,15 +248,17 @@ test.describe('processor detail workbench', () => {
   }) => {
     await installLangBotApiMocks(page, { authenticated: true });
     await page.route(
-      '**/api/v1/agents/agent-workbench/debug',
+      '**/api/v1/agents/agent-workbench/debug/stream',
       async (route) => {
         await route.fulfill({
-          status: 422,
-          contentType: 'application/json',
-          body: JSON.stringify({
-            code: 'dify.config_invalid',
-            msg: 'api-key is required',
-          }),
+          status: 200,
+          contentType: 'application/x-ndjson',
+          body:
+            JSON.stringify({
+              kind: 'error',
+              code: 'dify.config_invalid',
+              msg: 'api-key is required',
+            }) + '\n',
         });
       },
     );
@@ -454,8 +467,8 @@ test.describe('processor detail workbench', () => {
 
     const appShell = page.locator('[class*="group/sidebar-wrapper"]');
     const sidebarInset = page.locator('[data-slot="sidebar-inset"]');
-    await expect(appShell).toHaveCSS('overflow', 'clip');
-    await expect(sidebarInset).toHaveCSS('overflow', 'clip');
+    await expect(appShell).toHaveCSS('overflow', 'hidden');
+    await expect(sidebarInset).toHaveCSS('overflow', 'hidden');
     await expect
       .poll(() => appShell.evaluate((element) => element.scrollTop))
       .toBe(0);

web/tests/unit/agent-debug-execution.test.mjs

@@ -2,69 +2,137 @@ import assert from 'node:assert/strict';
 import fs from 'node:fs';
 import test from 'node:test';
 import ts from 'typescript';
-const source = fs.readFileSync(new URL('../../src/app/home/agents/components/debug-execution.ts', import.meta.url), 'utf8');
+const source = fs.readFileSync(
+  new URL(
+    '../../src/app/home/agents/components/debug-execution.ts',
+    import.meta.url,
+  ),
+  'utf8',
+);
 const module = { exports: {} };
-new Function('exports', ts.transpileModule(source, { compilerOptions: { module: ts.ModuleKind.CommonJS, target: ts.ScriptTarget.ES2022 } }).outputText)(module.exports);
+new Function(
+  'exports',
+  ts.transpileModule(source, {
+    compilerOptions: {
+      module: ts.ModuleKind.CommonJS,
+      target: ts.ScriptTarget.ES2022,
+    },
+  }).outputText,
+)(module.exports);
 const { executionSteps } = module.exports;
-const event = (type, data) => ({type, data});
+const event = (type, data) => ({ type, data });
 test('separates streamed thinking and text, replaces final snapshot without duplication', () => {
-  assert.deepEqual(executionSteps([
-    event('message.delta', {chunk: {content:'<think>plan'}}),
-    event('message.delta', {chunk: {content:'</think>hello'}}),
-    event('message.completed', {message: {content:'<think>plan</think>hello'}}),
-    event('run.completed', {message: {content:'hello'}}),
-  ]), [{kind:'message', text:'hello', reasoning:'plan'}]);
+  assert.deepEqual(
+    executionSteps([
+      event('message.delta', { chunk: { content: '<think>plan' } }),
+      event('message.delta', { chunk: { content: '</think>hello' } }),
+      event('message.completed', {
+        message: { content: '<think>plan</think>hello' },
+      }),
+      event('run.completed', { message: { content: 'hello' } }),
+    ]),
+    [{ kind: 'message', text: 'hello', reasoning: 'plan' }],
+  );
 });
 test('retains structured reasoning and tool parameters/results in order', () => {
   const steps = executionSteps([
-    event('message.delta', {chunk: {provider_specific_fields:{reasoning_content:'plan'}}}),
-    event('message.completed', {message: {content:''}}),
-    event('tool.call.started', {tool_call_id:'1',tool_name:'exec', parameters:{command:'echo hi'}}),
-    event('tool.call.started', {tool_call_id:'2',tool_name:'exec', parameters:{command:'bad'}}),
-    event('tool.call.completed', {tool_call_id:'2',tool_name:'exec', error:'failed'}),
-    event('tool.call.completed', {tool_call_id:'1',tool_name:'exec', result:{stdout:'hi'}}),
+    event('message.delta', {
+      chunk: { provider_specific_fields: { reasoning_content: 'plan' } },
+    }),
+    event('message.completed', { message: { content: '' } }),
+    event('tool.call.started', {
+      tool_call_id: '1',
+      tool_name: 'exec',
+      parameters: { command: 'echo hi' },
+    }),
+    event('tool.call.started', {
+      tool_call_id: '2',
+      tool_name: 'exec',
+      parameters: { command: 'bad' },
+    }),
+    event('tool.call.completed', {
+      tool_call_id: '2',
+      tool_name: 'exec',
+      error: 'failed',
+    }),
+    event('tool.call.completed', {
+      tool_call_id: '1',
+      tool_name: 'exec',
+      result: { stdout: 'hi' },
+    }),
     event('run.failed', {}),
   ]);
   assert.equal(steps[0].reasoning, 'plan');
   assert.equal(steps[1].parameters.command, 'echo hi');
-  assert.deepEqual(steps[1].result, {stdout:'hi'});
+  assert.deepEqual(steps[1].result, { stdout: 'hi' });
   assert.equal(steps[2].status, 'failed');
   assert.equal(steps[2].error, 'failed');
 });
 
 test('replaces LocalAgent cumulative snapshots instead of repeating text', () => {
-  assert.deepEqual(executionSteps([
-    event('message.delta', {chunk: {content:'hello', msg_sequence:1}}),
-    event('message.delta', {chunk: {content:'hello world', msg_sequence:2}}),
-    event('message.delta', {chunk: {content:'hello world', msg_sequence:3, is_final:true}}),
-  ]), [{kind:'message', text:'hello world', reasoning:''}]);
+  assert.deepEqual(
+    executionSteps([
+      event('message.delta', { chunk: { content: 'hello', msg_sequence: 1 } }),
+      event('message.delta', {
+        chunk: { content: 'hello world', msg_sequence: 2 },
+      }),
+      event('message.delta', {
+        chunk: { content: 'hello world', msg_sequence: 3, is_final: true },
+      }),
+    ]),
+    [{ kind: 'message', text: 'hello world', reasoning: '' }],
+  );
 });
 
 test('shows failed tool results even when the call transport completed', () => {
   const steps = executionSteps([
-    event('tool.call.started', {tool_call_id:'exit7', tool_name:'exec', parameters:{command:'exit 7'}}),
-    event('tool.call.completed', {tool_call_id:'exit7', tool_name:'exec', result:{ok:false, exit_code:7, stderr:'expected'}}),
+    event('tool.call.started', {
+      tool_call_id: 'exit7',
+      tool_name: 'exec',
+      parameters: { command: 'exit 7' },
+    }),
+    event('tool.call.completed', {
+      tool_call_id: 'exit7',
+      tool_name: 'exec',
+      result: { ok: false, exit_code: 7, stderr: 'expected' },
+    }),
   ]);
-  assert.equal(steps[0].status,'failed');
-  assert.equal(steps[0].result.exit_code,7);
+  assert.equal(steps[0].status, 'failed');
+  assert.equal(steps[0].result.exit_code, 7);
 });
 
 test('does not repeat prior thinking across LocalAgent tool turns', () => {
   const prefix = '<think>first thought</think>';
   const steps = executionSteps([
-    event('message.delta', {chunk:{content:prefix, msg_sequence:1}}),
-    event('tool.call.started', {tool_call_id:'w',tool_name:'write',parameters:{path:'/workspace/a'}}),
-    event('tool.call.completed', {tool_call_id:'w',tool_name:'write',result:{ok:true}}),
-    event('message.delta', {chunk:{content:prefix+'now read',msg_sequence:1}}),
-    event('tool.call.started', {tool_call_id:'r',tool_name:'read'}),
-    event('tool.call.completed', {tool_call_id:'r',tool_name:'read',result:{ok:true}}),
-    event('message.delta', {chunk:{content:prefix+'now read'+'done',msg_sequence:1}}),
-    event('message.completed', {message:{content:'done'}}),
+    event('message.delta', { chunk: { content: prefix, msg_sequence: 1 } }),
+    event('tool.call.started', {
+      tool_call_id: 'w',
+      tool_name: 'write',
+      parameters: { path: '/workspace/a' },
+    }),
+    event('tool.call.completed', {
+      tool_call_id: 'w',
+      tool_name: 'write',
+      result: { ok: true },
+    }),
+    event('message.delta', {
+      chunk: { content: prefix + 'now read', msg_sequence: 1 },
+    }),
+    event('tool.call.started', { tool_call_id: 'r', tool_name: 'read' }),
+    event('tool.call.completed', {
+      tool_call_id: 'r',
+      tool_name: 'read',
+      result: { ok: true },
+    }),
+    event('message.delta', {
+      chunk: { content: prefix + 'now read' + 'done', msg_sequence: 1 },
+    }),
+    event('message.completed', { message: { content: 'done' } }),
   ]);
-  const messages = steps.filter(s=>s.kind==='message');
+  const messages = steps.filter((s) => s.kind === 'message');
   assert.deepEqual(messages, [
-    {kind:'message',text:'',reasoning:'first thought'},
-    {kind:'message',text:'now read',reasoning:''},
-    {kind:'message',text:'done',reasoning:''},
+    { kind: 'message', text: '', reasoning: 'first thought' },
+    { kind: 'message', text: 'now read', reasoning: '' },
+    { kind: 'message', text: 'done', reasoning: '' },
   ]);
 });