Composio powers 1000+ toolkits, tool search, context management, authentication, and a sandboxed workbench to help you build AI agents that turn intent into action.
TypeScript agentic-aiagentsaiai-agentsaiagentsdeveloper-toolsfunction-callinggpt-4javascriptjsllmllmopsmcppythonremote-mcp-serverssetypescript
import { afterEach, vi } from 'vitest';
import { describe, expect, it, layer } from '@effect/vitest';
import { ConfigProvider, Effect } from 'effect';
import {
cliDebugFlagsLayer,
debugFlagsToChildEnv,
isAcpOnlyEnabled,
isPerfDebugEnabled,
isTelemetryDebugEnabled,
isToolDebugEnabled,
NO_CLI_DEBUG_FLAG_OVERRIDES,
stripTelemetryDebugFlag,
telemetryDebugModeLayer,
TELEMETRY_DEBUG_FLAG,
} from 'src/services/runtime-flags';
import { extendConfigProvider } from 'src/services/config';
import { TestLive } from 'test/__utils__';
describe('debugFlagsToChildEnv', () => {
it('[Given] resolved debug flags [Then] every flag serializes into the child environment', () => {
expect(
debugFlagsToChildEnv({
perfDebug: true,
toolDebug: false,
acpOnly: true,
telemetryDebug: false,
})
).toEqual({
COMPOSIO_PERF_DEBUG: '1',
COMPOSIO_TOOL_DEBUG: '0',
COMPOSIO_RUN_ACP_ONLY: '1',
COMPOSIO_CLI_TELEMETRY_DEBUG: '0',
});
});
});
describe('stripTelemetryDebugFlag', () => {
it('[Given] the flag before the delimiter [Then] it is removed and reported', () => {
expect(stripTelemetryDebugFlag(['bun', 'composio', TELEMETRY_DEBUG_FLAG, 'whoami'])).toEqual({
argv: ['bun', 'composio', 'whoami'],
telemetryDebug: true,
});
});
it('[Given] the flag after a `--` delimiter [Then] it belongs to the child and survives', () => {
const argv = ['bun', 'composio', 'run', 'my-agent', '--', TELEMETRY_DEBUG_FLAG];
expect(stripTelemetryDebugFlag(argv)).toEqual({ argv, telemetryDebug: false });
});
it('[Given] the flag on both sides of `--` [Then] only the CLI-side one is consumed', () => {
expect(
stripTelemetryDebugFlag([
'bun',
'composio',
TELEMETRY_DEBUG_FLAG,
'run',
'my-agent',
'--',
TELEMETRY_DEBUG_FLAG,
])
).toEqual({
argv: ['bun', 'composio', 'run', 'my-agent', '--', TELEMETRY_DEBUG_FLAG],
telemetryDebug: true,
});
});
it('[Given] no flag [Then] argv is reported unchanged', () => {
const argv = ['bun', 'composio', 'whoami'];
expect(stripTelemetryDebugFlag(argv)).toEqual({ argv, telemetryDebug: false });
});
});
const enabledDebugConfig = ConfigProvider.fromMap(
new Map([
['COMPOSIO_PERF_DEBUG', '1'],
['COMPOSIO_TOOL_DEBUG', '1'],
['COMPOSIO_RUN_ACP_ONLY', '1'],
])
).pipe(extendConfigProvider);
const readAllDebugFlags = Effect.all({
perfDebug: isPerfDebugEnabled,
toolDebug: isToolDebugEnabled,
acpOnly: isAcpOnlyEnabled,
});
describe('debug flag precedence', () => {
layer(TestLive())(it => {
it.effect('[Given] no flags [Then] the COMPOSIO_* config decides', () =>
Effect.gen(function* () {
expect(yield* readAllDebugFlags).toEqual({
perfDebug: true,
toolDebug: true,
acpOnly: true,
});
}).pipe(
Effect.provide(cliDebugFlagsLayer(NO_CLI_DEBUG_FLAG_OVERRIDES)),
Effect.withConfigProvider(enabledDebugConfig)
)
);
it.effect('[Given] explicit false flags [Then] they beat the enabled config', () =>
Effect.gen(function* () {
expect(yield* readAllDebugFlags).toEqual({
perfDebug: false,
toolDebug: false,
acpOnly: false,
});
}).pipe(
Effect.provide(cliDebugFlagsLayer({ perfDebug: false, toolDebug: false, acpOnly: false })),
Effect.withConfigProvider(enabledDebugConfig)
)
);
});
});
describe('telemetry debug mode', () => {
afterEach(() => {
vi.unstubAllEnvs();
});
it.effect('[Given] the bootstrap provided the flag [Then] it wins over the environment', () =>
Effect.gen(function* () {
vi.stubEnv('COMPOSIO_CLI_TELEMETRY_DEBUG', '');
expect(yield* isTelemetryDebugEnabled).toBe(true);
}).pipe(Effect.provide(telemetryDebugModeLayer(true)))
);
it.effect('[Given] no bootstrap flag [Then] it falls back to COMPOSIO_CLI_TELEMETRY_DEBUG', () =>
Effect.gen(function* () {
vi.stubEnv('COMPOSIO_CLI_TELEMETRY_DEBUG', 'true');
expect(yield* isTelemetryDebugEnabled).toBe(true);
vi.stubEnv('COMPOSIO_CLI_TELEMETRY_DEBUG', '');
expect(yield* isTelemetryDebugEnabled).toBe(false);
})
);
});