Consider it done. The open-source AI agent that works out of the box · 想到,就能做到。开源、开箱即用的 AI Agent。
TypeScript agentai-agentai-assistantandroidclaude-codecodexelectroniosllmmacosreact-nativetypescriptwindows
// @vitest-environment jsdom
import { act, renderHook, waitFor } from '@testing-library/react';
import { afterEach, describe, expect, it, vi } from 'vitest';
import type { SubagentRun, SubagentRunsListResponse } from '@cindy/maker-shared/subagent-workspace';
import { useSubagentRunStatusIndex } from '../useSubagentRunStatusIndex';
type ChangeListener = (
payload: { sessionId: string; runId: string | null },
stamp?: unknown,
) => void;
function run(partial: Partial<SubagentRun> & Pick<SubagentRun, 'id' | 'status'>): SubagentRun {
return {
parentSessionId: 's1',
provider: 'claude-code',
logicalAgentId: partial.id,
identityAliases: [],
providerRunIds: [],
capabilities: {} as SubagentRun['capabilities'],
startedAt: 1,
updatedAt: 1,
...partial,
};
}
function installApi(pages: Record<string, SubagentRunsListResponse>) {
const listeners = new Set<ChangeListener>();
const list = vi.fn(
async (input: { sessionId: string; cursor?: string }) =>
pages[`${input.sessionId}:${input.cursor ?? ''}`] ?? { supported: true, runs: [] },
);
const api = {
list,
onChanged: vi.fn((listener: ChangeListener) => {
listeners.add(listener);
return () => listeners.delete(listener);
}),
};
(window as unknown as { electronAPI: unknown }).electronAPI = {
localDb: { subagentRuns: api },
};
return { list, listeners };
}
afterEach(() => {
delete (window as unknown as { electronAPI?: unknown }).electronAPI;
});
describe('useSubagentRunStatusIndex', () => {
it('reads every page and indexes runs by their parent tool-use id', async () => {
installApi({
's1:': {
supported: true,
runs: [run({ id: 'r1', parentToolUseId: 'toolu_1', status: 'running' })],
nextCursor: 'c2',
},
's1:c2': {
supported: true,
runs: [run({ id: 'r2', parentToolUseId: 'toolu_2', status: 'failed' })],
},
});
const { result } = renderHook(() => useSubagentRunStatusIndex({ sessionId: 's1' }));
await waitFor(() => expect(result.current.get('toolu_2')).toBe('failed'));
expect(result.current.get('toolu_1')).toBe('running');
});
it('re-reads on a change push for its own task only', async () => {
const pages: Record<string, SubagentRunsListResponse> = {
's1:': {
supported: true,
runs: [run({ id: 'r1', parentToolUseId: 'toolu_1', status: 'running' })],
},
};
const { list, listeners } = installApi(pages);
const { result } = renderHook(() => useSubagentRunStatusIndex({ sessionId: 's1' }));
await waitFor(() => expect(result.current.get('toolu_1')).toBe('running'));
pages['s1:'] = {
supported: true,
runs: [run({ id: 'r1', parentToolUseId: 'toolu_1', status: 'completed', updatedAt: 2 })],
};
const callsBefore = list.mock.calls.length;
act(() => {
for (const listener of listeners) listener({ sessionId: 'other', runId: null });
});
act(() => {
for (const listener of listeners) listener({ sessionId: 's1', runId: 'r1' });
});
await waitFor(() => expect(result.current.get('toolu_1')).toBe('completed'));
expect(list.mock.calls.length).toBe(callsBefore + 1);
});
it('falls back to the live status when a later refresh fails', async () => {
const { list, listeners } = installApi({
's1:': {
supported: true,
runs: [run({ id: 'r1', parentToolUseId: 'toolu_1', status: 'completed' })],
},
});
const { result } = renderHook(() => useSubagentRunStatusIndex({ sessionId: 's1' }));
await waitFor(() => expect(result.current.get('toolu_1')).toBe('completed'));
list.mockRejectedValueOnce(new Error('db busy'));
act(() => {
for (const listener of listeners) listener({ sessionId: 's1', runId: 'r1' });
});
await waitFor(() => expect(result.current.size).toBe(0));
});
it('does not read this machine for a device-link task', () => {
const { list } = installApi({});
const { result } = renderHook(() =>
useSubagentRunStatusIndex({ sessionId: 's1', deviceId: 'device-2' }),
);
expect(list).not.toHaveBeenCalled();
expect(result.current.size).toBe(0);
});
it('never shows the previous task while the next task is still loading', async () => {
installApi({
's1:': {
supported: true,
runs: [run({ id: 'r1', parentToolUseId: 'toolu_1', status: 'running' })],
},
});
const { result, rerender } = renderHook(
({ sessionId }) => useSubagentRunStatusIndex({ sessionId }),
{ initialProps: { sessionId: 's1' } },
);
await waitFor(() => expect(result.current.get('toolu_1')).toBe('running'));
rerender({ sessionId: 's2' });
expect(result.current.get('toolu_1')).toBeUndefined();
// Let the s2 read settle inside act so it cannot leak into the next test.
await waitFor(() => expect(result.current.size).toBe(0));
});
});