The fastest, cheapest, most efficient open-source agent harness. Run more than 500 agents on a $10 VPS.
Rust agent-orchestrationai-agentsai-assistantdesktopllmlocal-firstmcppersonal-aiprivacyrustsecond-braintauri
import { fireEvent, render, screen, within } from '@testing-library/react';
import { describe, expect, it } from 'vitest';
import type { SubagentActivity } from '../../../store/chatRuntimeSlice';
import { SubagentActivityCard } from './SubagentActivityCard';
function openDisclosure() {
fireEvent.click(within(screen.getByTestId('assistant-ui-subagent-call')).getByRole('button'));
}
describe('SubagentActivityCard', () => {
it('derives a working state from a running activity with no explicit status text needed', () => {
render(
<SubagentActivityCard
activity={{ taskId: 't', agentId: 'researcher', status: 'running', toolCalls: [] }}
/>
);
const card = screen.getByTestId('assistant-ui-subagent-call');
expect(card).toHaveAttribute('data-state', 'working');
expect(card).toHaveAttribute('data-status', 'running');
});
it('marks a failed delegation as failed rather than complete', () => {
render(
<SubagentActivityCard
activity={{ taskId: 't', agentId: 'researcher', status: 'failed', toolCalls: [] }}
/>
);
const card = screen.getByTestId('assistant-ui-subagent-call');
expect(card).toHaveAttribute('data-state', 'failed');
expect(card).toHaveAttribute('data-status', 'failed');
});
it('shows an incomplete delegation as a partial state with a status announcement', () => {
render(
<SubagentActivityCard
activity={{ taskId: 't', agentId: 'researcher', status: 'incomplete', toolCalls: [] }}
/>
);
const card = screen.getByTestId('assistant-ui-subagent-call');
expect(card).toHaveAttribute('data-state', 'incomplete');
expect(screen.getByTestId('subagent-incomplete')).toHaveAttribute('role', 'status');
});
it('marks a cancelled delegation as cancelled', () => {
render(
<SubagentActivityCard
activity={{ taskId: 't', agentId: 'researcher', status: 'cancelled', toolCalls: [] }}
/>
);
expect(screen.getByTestId('assistant-ui-subagent-call')).toHaveAttribute(
'data-state',
'cancelled'
);
});
it('shows the awaiting-user question as plain text with no reply box', () => {
// This surface (the inline rail, the Agent Process Source panel) renders
// outside an AssistantRuntimeProvider, so unlike `SubagentTaskCard` there
// is nowhere to send a reply through — the question is read-only here.
render(
<SubagentActivityCard
activity={{
taskId: 't',
agentId: 'researcher',
status: 'awaiting_user',
awaitingQuestion: 'Which repo should I use?',
toolCalls: [],
}}
/>
);
expect(screen.getByTestId('assistant-ui-subagent-call')).toHaveAttribute(
'data-status',
'awaiting_user'
);
expect(screen.getByTestId('subagent-awaiting-question')).toHaveTextContent(
'Which repo should I use?'
);
expect(screen.queryByTestId('subagent-answer-input')).toBeNull();
});
it('renders the delegation label with the agent display name', () => {
render(
<SubagentActivityCard
activity={{
taskId: 't',
agentId: 'researcher',
displayName: 'Researcher',
status: 'success',
toolCalls: [],
}}
/>
);
expect(screen.getByText('Delegated to Researcher')).toBeInTheDocument();
});
it('keeps completed output and ignores legacy worktree fields', () => {
const legacyActivity: SubagentActivity & {
worktreePath: string;
changedFiles: string[];
isDirty: boolean;
} = {
taskId: 't',
agentId: 'researcher',
status: 'success',
output: 'Research is complete.',
toolCalls: [],
worktreePath: '/repo/.claude/worktrees/worker',
changedFiles: ['notes.md'],
isDirty: true,
};
render(<SubagentActivityCard activity={legacyActivity} />);
expect(screen.getByText('Research is complete.')).toBeInTheDocument();
expect(screen.queryByTestId('worktree-actions')).toBeNull();
expect(screen.queryByText(/worker/)).toBeNull();
});
it('renders child tool calls (from `toolCalls`) inside the nested transcript once opened', () => {
render(
<SubagentActivityCard
activity={{
taskId: 't',
agentId: 'researcher',
status: 'success',
toolCalls: [{ callId: 'c1', toolName: 'web_search', status: 'success' }],
}}
/>
);
openDisclosure();
expect(screen.getByTestId('subagent-activity')).toBeInTheDocument();
});
it('renders no nested-transcript disclosure chevron when there is nothing to show', () => {
render(
<SubagentActivityCard activity={{ taskId: 't', agentId: 'researcher', toolCalls: [] }} />
);
expect(screen.getByRole('button', { name: /Delegated to researcher/i })).toHaveAttribute(
'disabled'
);
});
});