feat(client): resume + getScreen on the automation SDK surface (#3528)

zhuangjianguo · claude · zhuangjianguo · commit 2b4b34b0dd82 · 2026-07-27T04:16:57.000Z
A `type: 'screen'` flow suspends at its `screen` node — `execute()` returns `{ status: 'paused', runId, screen }` and the run waits for input. The dispatcher has served the other half of that contract since ADR-0019 (`POST /automation/:flow/runs/:runId/resume`, `GET .../screen`), but the client SDK's automation surface stopped at getFlow / execute / listRuns / getRun. Any consumer built on the SDK could therefore start a screen flow and never finish it: the run stayed suspended and the only way out was hand-rolling the HTTP call. That is what dead-ended the Console's developer "Flow Runs" test runner, where every test run of a screen flow orphaned a `paused` row. - `automation.resume(flowName, runId, signal?)` posts the collected screen values as `inputs`, plus the approval-style `output` / `branchLabel` the dispatcher already accepts, and returns either the next paused screen of a multi-step wizard or the terminal AutomationResult. - `automation.getScreen(flowName, runId)` returns the screen a paused run is waiting on, so a client that did not launch the run (a reload, another tab, an inbox) can render the pending step before resuming. - Both land on the environment-scoped client as well as the unscoped one. The two dispatcher routes had no test coverage at all; they now cover the `inputs` / `variables` aliasing, approval-style output + branchLabel, the pause-again envelope, the 501 when the service cannot resume, and the ordering guard that keeps `/runs/:runId/screen` from being swallowed by the run lookup. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LX9ut3MK3KykE11S9bJmv5
diff --git a/.changeset/automation-client-resume-screen-flow.md b/.changeset/automation-client-resume-screen-flow.md
@@ -0,0 +1,30 @@
+---
+"@objectstack/client": minor
+---
+
+feat(client): `automation.resume()` / `automation.getScreen()` — finish a paused screen flow from the SDK (#3528)
+
+A `type: 'screen'` flow suspends when it reaches a `screen` node: `execute()`
+returns `{ status: 'paused', runId, screen }` and the run waits for input. The
+second half of that contract — `POST /automation/:flow/runs/:runId/resume` —
+has shipped in the dispatcher since ADR-0019, but the client SDK's automation
+surface stopped at `getFlow` / `execute` / `listRuns` / `getRun`. Anything built
+on the SDK could therefore *start* a screen flow and never finish it: the run
+stayed suspended and the only way out was hand-rolling the HTTP call. That gap
+is what stranded the Console's developer "Flow Runs" test runner, where every
+test run of a screen flow orphaned a `paused` row.
+
+- **`automation.resume(flowName, runId, signal?)`** — posts the collected screen
+  values as `inputs` (applied as bare flow variables), plus the approval-style
+  `output` / `branchLabel` the dispatcher already accepts. Returns the next
+  `{ status: 'paused', screen }` of a multi-step wizard, or the terminal
+  `AutomationResult`.
+- **`automation.getScreen(flowName, runId)`** — the screen a paused run is
+  waiting on, so a client that did not launch the run (a page reload, another
+  tab, an inbox) can render the pending step before resuming.
+- Both are available on the environment-scoped client
+  (`client.project(id).automation.*`) as well as the unscoped one.
+
+Also covers the two dispatcher routes with tests — the resume and screen paths
+had none, including the ordering guard that keeps `/runs/:runId/screen` from
+being swallowed by the `/runs/:runId` run lookup.
diff --git a/packages/client/src/client.test.ts b/packages/client/src/client.test.ts
@@ -852,6 +852,83 @@ describe('ObjectStackClient.automation', () => {
         );
     });
 
+    // ── screen-flow runtime (ADR-0019 durable pause, #3528) ──────────────
+
+    it('should resume a paused run with the collected screen input', async () => {
+        const { client, fetchMock } = createMockClient({
+            success: true,
+            data: { success: true, output: {}, durationMs: 12 },
+        });
+
+        const result = await client.automation.resume('my_flow', 'run_1', {
+            inputs: { new_assignee: 'ada@example.com' },
+        });
+        expect(fetchMock).toHaveBeenCalledWith(
+            'http://localhost:3000/api/v1/automation/my_flow/runs/run_1/resume',
+            expect.objectContaining({
+                method: 'POST',
+                body: JSON.stringify({ inputs: { new_assignee: 'ada@example.com' } }),
+            }),
+        );
+        expect(result.success).toBe(true);
+    });
+
+    it('should resume with an approval branch label and node output', async () => {
+        const { client, fetchMock } = createMockClient({ success: true, data: { success: true } });
+
+        await client.automation.resume('my_flow', 'run_1', {
+            output: { comment: 'looks good' },
+            branchLabel: 'approve',
+        });
+        expect(fetchMock).toHaveBeenCalledWith(
+            'http://localhost:3000/api/v1/automation/my_flow/runs/run_1/resume',
+            expect.objectContaining({
+                method: 'POST',
+                body: JSON.stringify({ output: { comment: 'looks good' }, branchLabel: 'approve' }),
+            }),
+        );
+    });
+
+    it('should post an empty signal when resuming with no input', async () => {
+        const { client, fetchMock } = createMockClient({ success: true, data: { success: true } });
+
+        await client.automation.resume('my flow', 'run 1');
+        expect(fetchMock).toHaveBeenCalledWith(
+            'http://localhost:3000/api/v1/automation/my%20flow/runs/run%201/resume',
+            expect.objectContaining({ method: 'POST', body: '{}' }),
+        );
+    });
+
+    it('should return the next screen when a multi-step wizard pauses again', async () => {
+        const { client } = createMockClient({
+            success: true,
+            data: {
+                success: true,
+                status: 'paused',
+                runId: 'run_1',
+                screen: { nodeId: 'step2', title: 'Opportunity', fields: [] },
+            },
+        });
+
+        const result = await client.automation.resume('my_flow', 'run_1', { inputs: { account_id: 'a1' } });
+        expect(result.status).toBe('paused');
+        expect(result.screen.nodeId).toBe('step2');
+    });
+
+    it('should fetch the screen a paused run awaits', async () => {
+        const { client, fetchMock } = createMockClient({
+            success: true,
+            data: { runId: 'run_1', screen: { nodeId: 'collect', fields: [] } },
+        });
+
+        const result = await client.automation.getScreen('my_flow', 'run_1');
+        expect(fetchMock).toHaveBeenCalledWith(
+            'http://localhost:3000/api/v1/automation/my_flow/runs/run_1/screen',
+            expect.any(Object),
+        );
+        expect(result.screen.nodeId).toBe('collect');
+    });
+
     // ==========================================
     // capabilities getter
     // ==========================================
@@ -1062,6 +1139,23 @@ describe('ScopedProjectClient', () => {
         const scoped = client.project('00000000-0000-0000-0000-000000000001');
         expect(scoped.getProjectId()).toBe('00000000-0000-0000-0000-000000000001');
     });
+
+    it('prefixes the screen-flow automation.resume / getScreen calls', async () => {
+        const { client, fetchMock } = createMockClient({ success: true, data: { success: true } });
+        const scoped = client.project('proj-123');
+
+        await scoped.automation.resume('my_flow', 'run_1', { inputs: { note: 'ok' } });
+        expect(fetchMock).toHaveBeenLastCalledWith(
+            'http://localhost:3000/api/v1/environments/proj-123/automation/my_flow/runs/run_1/resume',
+            expect.objectContaining({ method: 'POST', body: JSON.stringify({ inputs: { note: 'ok' } }) }),
+        );
+
+        await scoped.automation.getScreen('my_flow', 'run_1');
+        expect(fetchMock).toHaveBeenLastCalledWith(
+            'http://localhost:3000/api/v1/environments/proj-123/automation/my_flow/runs/run_1/screen',
+            expect.any(Object),
+        );
+    });
 });
 
 // ==========================================
diff --git a/packages/client/src/index.ts b/packages/client/src/index.ts
@@ -2311,6 +2311,48 @@ export class ObjectStackClient {
           );
           return this.unwrapResponse(res) as Promise<T>;
       },
+      /**
+       * Resume a run suspended at a `screen` (or `approval`) node — the
+       * screen-flow runtime's second half (ADR-0019 durable pause).
+       *
+       * `execute()` returns `{ status: 'paused', runId, screen }` when a flow
+       * reaches a screen node; the collected values go back through here as
+       * `inputs` (applied as bare flow variables). The result is either the
+       * NEXT `{ status: 'paused', screen }` of a multi-step wizard or the
+       * terminal `AutomationResult`. Without this method a paused run can only
+       * be finished by hand-rolling the HTTP call (#3528).
+       */
+      resume: async <T = any>(
+          flowName: string,
+          runId: string,
+          signal?: {
+              /** Screen input values, applied as bare flow variables. */
+              inputs?: Record<string, unknown>;
+              /** Node output, namespaced under the suspended node's id. */
+              output?: Record<string, unknown>;
+              /** Out-edge to follow (e.g. an approval's `approve` / `reject`). */
+              branchLabel?: string;
+          },
+      ): Promise<T> => {
+          const route = this.getRoute('automation');
+          const res = await this.fetch(
+              `${this.baseUrl}${route}/${encodeURIComponent(flowName)}/runs/${encodeURIComponent(runId)}/resume`,
+              { method: 'POST', body: JSON.stringify(signal ?? {}) },
+          );
+          return this.unwrapResponse(res) as Promise<T>;
+      },
+      /**
+       * Fetch the screen a paused run is waiting on — lets a client that did
+       * not launch the run (a reload, a different tab, an inbox) render the
+       * pending step before calling {@link resume}.
+       */
+      getScreen: async <T = any>(flowName: string, runId: string): Promise<T> => {
+          const route = this.getRoute('automation');
+          const res = await this.fetch(
+              `${this.baseUrl}${route}/${encodeURIComponent(flowName)}/runs/${encodeURIComponent(runId)}/screen`,
+          );
+          return this.unwrapResponse(res) as Promise<T>;
+      },
   };
 
   /**
@@ -3644,6 +3686,33 @@ export class ScopedProjectClient {
       );
       return this.parent._unwrap<T>(res);
     },
+    /**
+     * Resume a run suspended at a `screen` / `approval` node with the collected
+     * input (ADR-0019 durable pause). Mirrors the unscoped
+     * `client.automation.resume`.
+     */
+    resume: async <T = any>(
+      flowName: string,
+      runId: string,
+      signal?: {
+        inputs?: Record<string, unknown>;
+        output?: Record<string, unknown>;
+        branchLabel?: string;
+      },
+    ): Promise<T> => {
+      const res = await this.parent._fetch(
+        this.url(`/automation/${encodeURIComponent(flowName)}/runs/${encodeURIComponent(runId)}/resume`),
+        { method: 'POST', body: JSON.stringify(signal ?? {}) },
+      );
+      return this.parent._unwrap<T>(res);
+    },
+    /** Fetch the screen a paused run is waiting on. */
+    getScreen: async <T = any>(flowName: string, runId: string): Promise<T> => {
+      const res = await this.parent._fetch(
+        this.url(`/automation/${encodeURIComponent(flowName)}/runs/${encodeURIComponent(runId)}/screen`),
+      );
+      return this.parent._unwrap<T>(res);
+    },
   };
 }
 
diff --git a/packages/runtime/src/http-dispatcher.test.ts b/packages/runtime/src/http-dispatcher.test.ts
@@ -168,6 +168,9 @@ describe('HttpDispatcher', () => {
                 listRuns: vi.fn().mockResolvedValue([{ id: 'run_1', status: 'completed' }]),
                 getRun: vi.fn().mockResolvedValue({ id: 'run_1', status: 'completed' }),
                 trigger: vi.fn().mockResolvedValue({ success: true }),
+                resume: vi.fn().mockResolvedValue({ success: true, output: {}, durationMs: 7 }),
+                // Sync per IAutomationService — `ScreenSpec | null`, not a promise.
+                getSuspendedScreen: vi.fn().mockReturnValue({ nodeId: 'collect', fields: [] }),
                 getActionDescriptors: vi.fn().mockReturnValue([
                     { type: 'decision', name: 'Decision', category: 'logic', paradigms: ['flow'], source: 'builtin' },
                     { type: 'http_request', name: 'HTTP Request', category: 'io', paradigms: ['flow', 'approval'], source: 'builtin' },
@@ -275,6 +278,80 @@ describe('HttpDispatcher', () => {
             expect(result.response?.status).toBe(404);
         });
 
+        // ── screen-flow runtime (ADR-0019 durable pause, #3528) ──────────
+        it('should resume a paused run via POST /:name/runs/:runId/resume', async () => {
+            const result = await dispatcher.handleAutomation(
+                'flow_a/runs/run_1/resume', 'POST', { inputs: { new_assignee: 'ada' } }, { request: {} },
+            );
+            expect(result.handled).toBe(true);
+            expect(mockAutomationService.resume).toHaveBeenCalledWith('run_1', {
+                variables: { new_assignee: 'ada' },
+            });
+            expect(result.response?.body?.data?.success).toBe(true);
+        });
+
+        it('should accept `variables` as an alias for `inputs` on resume', async () => {
+            await dispatcher.handleAutomation(
+                'flow_a/runs/run_1/resume', 'POST', { variables: { note: 'hi' } }, { request: {} },
+            );
+            expect(mockAutomationService.resume).toHaveBeenCalledWith('run_1', {
+                variables: { note: 'hi' },
+            });
+        });
+
+        it('should forward approval-style output + branchLabel on resume', async () => {
+            await dispatcher.handleAutomation(
+                'flow_a/runs/run_1/resume', 'POST',
+                { output: { comment: 'ok' }, branchLabel: 'approve' }, { request: {} },
+            );
+            expect(mockAutomationService.resume).toHaveBeenCalledWith('run_1', {
+                output: { comment: 'ok' },
+                branchLabel: 'approve',
+            });
+        });
+
+        it('should resume with an empty signal when the body carries no input', async () => {
+            await dispatcher.handleAutomation('flow_a/runs/run_1/resume', 'POST', undefined, { request: {} });
+            expect(mockAutomationService.resume).toHaveBeenCalledWith('run_1', {});
+        });
+
+        it('should surface the next screen when a resumed run pauses again', async () => {
+            mockAutomationService.resume.mockResolvedValue({
+                success: true, status: 'paused', runId: 'run_1',
+                screen: { nodeId: 'step2', title: 'Confirm', fields: [] },
+            });
+            const result = await dispatcher.handleAutomation(
+                'flow_a/runs/run_1/resume', 'POST', { inputs: {} }, { request: {} },
+            );
+            expect(result.response?.body?.data?.status).toBe('paused');
+            expect(result.response?.body?.data?.screen?.nodeId).toBe('step2');
+        });
+
+        it('should return 501 when the automation service cannot resume', async () => {
+            delete mockAutomationService.resume;
+            const result = await dispatcher.handleAutomation(
+                'flow_a/runs/run_1/resume', 'POST', { inputs: {} }, { request: {} },
+            );
+            expect(result.handled).toBe(true);
+            expect(result.response?.status).toBe(501);
+        });
+
+        it('should get the pending screen via GET /:name/runs/:runId/screen', async () => {
+            const result = await dispatcher.handleAutomation('flow_a/runs/run_1/screen', 'GET', {}, { request: {} });
+            expect(result.handled).toBe(true);
+            expect(mockAutomationService.getSuspendedScreen).toHaveBeenCalledWith('run_1');
+            expect(result.response?.body?.data?.screen?.nodeId).toBe('collect');
+            // `screen` must NOT be swallowed by the getRun route below it.
+            expect(mockAutomationService.getRun).not.toHaveBeenCalled();
+        });
+
+        it('should return 404 when the run is not awaiting a screen', async () => {
+            mockAutomationService.getSuspendedScreen.mockReturnValue(null);
+            const result = await dispatcher.handleAutomation('flow_a/runs/run_1/screen', 'GET', {}, { request: {} });
+            expect(result.handled).toBe(true);
+            expect(result.response?.status).toBe(404);
+        });
+
         it('should handle legacy trigger path POST /trigger/:name', async () => {
             const result = await dispatcher.handleAutomation('trigger/flow_a', 'POST', { data: 1 }, { request: {} });
             expect(result.handled).toBe(true);
