Evenfire is an intelligence layer platform for running LLM agents that take real actions, on your infrastructure
TypeScript
import { defineConfig } from 'vitest/config'
export default defineConfig({
test: {
include: ['**/*.test.ts', 'integration/**/*.test.ts'],
exclude: ['**/node_modules/**', '**/fixtures/**/node_modules/**'],
testTimeout: 360_000, // 360s — workflow lifecycle tests wait up to 5 min for LLM execution
hookTimeout: 360_000,
fileParallelism: false,
sequence: { concurrent: false },
reporters: ['verbose'],
// A run that matched zero test files must FAIL, never exit 0. Vitest's
// default is `true` (fail-open); an e2e gate that resolves an empty glob
// proves nothing. The runner (scripts/e2e/run-vitest-e2e.sh) adds a
// per-suite existence precheck and a "Tests N>0" floor guard on top.
passWithNoTests: false,
},
})