Open-source, secure environment with real-world tools for enterprise-grade agents.
Python agentaiai-agentai-agentscode-interpretercopilotdevelopmentdevtoolsgptgpt-4javascriptllmnextjsopenaipythonreactsoftwaretypescript
import { vi } from 'vitest'
// msw mocks the network by patching `globalThis.fetch`, which the undici
// dispatcher the SDK uses on Node bypasses entirely — mocked requests would
// escape to the real API. Registered via `setupFiles` for every Node project,
// this reroutes the SDK through its real undici-unavailable fallback, which
// late-binds `globalThis.fetch`, so msw suites need no per-file mock. Tests
// that inject their own `loadUndici` (the api/envd transport suites) keep it,
// so the dispatcher wiring itself stays covered.
vi.mock('../src/undici', async (importOriginal) => {
const original = await importOriginal<typeof import('../src/undici')>()
return {
...original,
buildDispatchedFetch: (
options: Parameters<typeof original.buildDispatchedFetch>[0]
) =>
original.buildDispatchedFetch({
...options,
loadUndici: options.loadUndici ?? (async () => undefined),
}),
}
})