Evenfire is an intelligence layer platform for running LLM agents that take real actions, on your infrastructure
TypeScript
import { beforeAll, vi } from 'vitest'
import type { CommunicationChannelCRD, Message } from '../src/types'
export let ChannelReader: typeof import('../src/main').ChannelReader
beforeAll(async () => {
process.env.CLERUM_DEV_MODE = 'true'
;({ ChannelReader } = await import('../src/main'))
})
export function makeMessage(content: string): Message {
return {
channelType: 'telegram',
channelId: 'telegram-chat-1',
sender: '123456',
content,
timestamp: new Date('2026-05-04T00:00:00.000Z'),
messageId: `msg-${content.replace(/\W+/g, '-')}`,
}
}
export function makeChannel(): CommunicationChannelCRD {
return {
name: 'shared-telegram',
namespace: 'channels',
spec: {
hostRef: 'chatllm',
telegram: [{ channelId: 'telegram-chat-1', userIds: ['123456'] }],
},
}
}
export function makeRpcClient(overrides: Record<string, unknown> = {}) {
return {
healthCheck: vi.fn(async () => true),
sendMessage: vi.fn(),
getBaseUrl: vi.fn(() => 'http://shared-mcp-host.test'),
getTaskResult: vi.fn(),
sendApproval: vi.fn(),
sendDenial: vi.fn(),
sendWorkflowApprovalDecision: vi.fn(async () => ({ success: true, duplicate: false })),
resolveWorkflowApproval: vi.fn(async () => null),
getCronResults: vi.fn(async () => []),
acknowledgeCronResult: vi.fn(),
...overrides,
}
}