Self-hosted, Self-Improving, Federated Everything App — dev management, agent orchestration, persistent mind, digital twin, second brain, and creative pipeline from a single dashboard over Tailscale to port the user into the OS
JavaScript ai-agentsdashboarddeveloper-toolsdigital-twinquantified-selfsecond-brainself-hostedtailscale
import { describe, it, expect, vi, beforeEach } from 'vitest';
import express from 'express';
import { request } from '../../lib/testHelper.js';
import agentRoutes from './agents.js';
vi.mock('../../services/apps.js', () => ({
getAppById: vi.fn(),
}));
vi.mock('../../services/cosAgentLifecycle.js', () => ({
getAgents: vi.fn(),
}));
vi.mock('../../services/cosAgentIndex.js', () => ({
getAgentDates: vi.fn(),
getAgentsByDate: vi.fn(),
}));
import * as appsService from '../../services/apps.js';
import { getAgentDates, getAgentsByDate } from '../../services/cosAgentIndex.js';
import { getAgents } from '../../services/cosAgentLifecycle.js';
const APP = { id: 'app-001', name: 'Widget' };
const agentFor = (id, overrides = {}) => ({
id,
status: 'completed',
output: [{ line: 'a transcript line' }],
metadata: { taskApp: APP.id, taskDescription: 'Fix the widget' },
...overrides,
});
describe('GET /api/apps/:id/agents', () => {
let app;
beforeEach(() => {
app = express();
app.use('/api/apps', agentRoutes);
vi.clearAllMocks();
appsService.getAppById.mockResolvedValue(APP);
getAgents.mockResolvedValue([]);
getAgentDates.mockResolvedValue([]);
getAgentsByDate.mockResolvedValue([]);
});
// This response carries up to 500 records; shipping each one's transcript and
// whole pasted-prompt description is the same payload problem the CoS Agents
// tab had. The table here renders the description as a clamped link label and
// never reuses the text, so the bounded copy is all it needs.
it('projects the combined list: no transcripts, bounded descriptions', async () => {
const long = 'Refactor the queue. '.repeat(5000);
getAgents.mockResolvedValue([agentFor('agent-running', { status: 'running' })]);
getAgentDates.mockResolvedValue([{ date: '2026-07-13', count: 1 }]);
getAgentsByDate.mockResolvedValue([
agentFor('agent-done', { metadata: { taskApp: APP.id, taskDescription: long } }),
]);
const response = await request(app).get('/api/apps/app-001/agents');
expect(response.status).toBe(200);
expect(response.body.agents).toHaveLength(2);
expect(response.body.agents.every(a => !('output' in a))).toBe(true);
const archived = response.body.agents.find(a => a.id === 'agent-done');
expect(archived.metadata.taskDescription.length).toBeLessThan(long.length);
expect(archived.metadata.taskDescriptionTruncated).toBe(true);
// The summary counts still read the real statuses, not the projection.
expect(response.body.summary).toMatchObject({ total: 2, running: 1, succeeded: 1 });
});
});