ktx/packages/cli/test/llm/model-provider.test.ts
Andrey Avtomonov 494618ab14
feat: add codex llm backend for ktx runtime work (#253)
* feat: add codex sdk runner foundation

* feat: parse codex runtime events

* feat: expose codex runtime mcp tools

* feat: add codex llm runtime

* feat: wire codex llm backend

* test: avoid Array.fromAsync in codex runner test

* docs: document codex llm backend

* fix: tighten codex runtime config ownership

* fix: use codex sdk env and thread options

* fix: parse codex sdk event shapes

* test: add codex backend live smoke

* docs: clarify codex backend isolation

* fix: drive codex loop metrics from mcp events

* fix: enforce codex local step budget

* docs: disclose codex isolation limits

* fix: count all codex agent steps and stream step callbacks live

The agent-loop step budget only counted completed mcp_tool_call items, so
built-in command_execution steps (which the public Codex SDK/CLI surface can
still expose) never decremented the budget, letting ingest/reconciliation run
past stepBudget until Codex stopped on its own. onStepFinish was also replayed
only after the whole stream drained, so live work_unit_step / reconciliation
progress appeared stuck until the Codex process exited.

collectEvents is now the single live step accumulator: it counts every
completed agent-action item via a shared isCompletedAgentStep predicate
(command_execution, mcp_tool_call, file_change, web_search), fires onStepFinish
as each step completes, and enforces the budget on that broader count. A
no-tool turn still counts as one step. toolFailures stays MCP-specific, since a
non-zero command exit is normal agent exploration, not a loop failure.

* test: align ingest llm-guard assertions with codex backend

The skip-llm ingest guard message now lists codex as a valid backend and
mentions a Claude Code/Codex session plus a codex setup hint, but this slow
suite test still asserted the pre-codex wording. Update it to match the
production message (already covered by the local-bundle-runtime unit test) and
add the codex setup-line assertion.

* fix: treat codex error:null tool calls as success

The Codex SDK serializes error: null on successful mcp_tool_call items, so
the failure check (item.error !== undefined) flagged every successful tool
call as failed with the empty-payload default "Codex turn failed". This
killed every ingest work unit under the codex backend before it could
produce a patch.

Key on status === 'failed' (authoritative, always set) and only treat a
populated error object as a failure. Add a regression test built from a
verbatim real-SDK event capture.

* fix: default codex backend to gpt-5.5 and report real probe errors

The previous default gpt-5.3-codex is an API-key-only model that the OpenAI
API rejects under ChatGPT-account (subscription) auth, so codex status/setup
failed with a misleading "authentication is not usable" message even though
auth was fine.

- Default codex model is now gpt-5.5 (works on both subscription and API-key
  auth); the curated setup picker offers gpt-5.5 / gpt-5.4 / gpt-5.4-mini and
  keeps free-form entry for account-specific ids (e.g. gpt-5.3-codex-spark).
- runCodexAuthProbe now distinguishes "model not available" from an auth
  failure and surfaces the real API error: collectEvents retains stream
  events when the SDK throws on a non-zero exit, and the API error JSON
  envelope is unwrapped to its human-readable message.
- The Codex isolation warning now renders inside the clack setup frame.
- Docs updated to gpt-5.5 with a note that *-codex ids require API-key auth.

* fix: require llm.models.default in status and match codex probe remediation

Status reported a project ready when a non-none LLM backend was configured
without llm.models.default, but the runtime (resolveModelSlots) hard-requires
it, so ingest/scan/memory threw after `ktx status` said the project was usable.
buildLlmStatus now fails for any non-none backend missing models.default and no
longer invents a fallback model for claude-code/codex.

Codex probe failures now carry a category-matched fix: a model-access failure
steers the user at llm.models.default instead of the auth/install remediation.
runCodexAuthProbe returns the fix and status consumes it; the message stays
self-sufficient so setup output is unchanged.

Docs: README now lists the codex backend and local Codex auth; ktx-setup.mdx
states --llm-model only accepts codex/default or gpt-*/codex-* ids.

Repaired four doctor fixtures that configured a backend without models.default
(the now-correctly-blocked config) and added coverage for the new behavior.
2026-06-02 13:57:11 +02:00

324 lines
12 KiB
TypeScript

import { devToolsMiddleware as defaultDevToolsMiddleware } from '@ai-sdk/devtools';
import { wrapLanguageModel as defaultWrapLanguageModel, type LanguageModel } from 'ai';
import { describe, expect, it, vi } from 'vitest';
import { createKtxLlmProvider, type KtxLlmProviderFactoryDeps } from '../../src/llm/model-provider.js';
const languageModel = (modelId: string, provider = 'test'): LanguageModel => ({ modelId, provider }) as LanguageModel;
const devtoolsMiddleware = (): ReturnType<typeof defaultDevToolsMiddleware> => ({ specificationVersion: 'v3' });
const wrapWith = (model: LanguageModel) =>
vi.fn((_options: Parameters<typeof defaultWrapLanguageModel>[0]) => model as ReturnType<typeof defaultWrapLanguageModel>);
describe('createKtxLlmProvider', () => {
it('wraps language models with DevTools middleware when explicitly enabled', () => {
const anthropicModel = languageModel('claude-sonnet-4-6', 'anthropic');
const wrappedModel = languageModel('claude-sonnet-4-6', 'anthropic-devtools');
const middleware = devtoolsMiddleware();
const wrapLanguageModel = wrapWith(wrappedModel);
const devToolsMiddleware = vi.fn(devtoolsMiddleware);
const provider = createKtxLlmProvider(
{
backend: 'anthropic',
anthropic: { apiKey: 'test-anthropic-key' }, // pragma: allowlist secret
modelSlots: { default: 'claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{
createAnthropic: vi.fn(() => vi.fn(() => anthropicModel)),
devtoolsEnabled: true,
wrapLanguageModel,
devToolsMiddleware,
} satisfies KtxLlmProviderFactoryDeps,
);
expect(provider.getModel('default')).toBe(wrappedModel);
expect(devToolsMiddleware).toHaveBeenCalledTimes(1);
expect(wrapLanguageModel).toHaveBeenCalledWith({
model: anthropicModel,
middleware,
modelId: 'claude-sonnet-4-6',
providerId: 'anthropic',
});
});
it('does not wrap language models by default', () => {
const anthropicModel = languageModel('claude-sonnet-4-6', 'anthropic');
const wrapLanguageModel = vi.fn(defaultWrapLanguageModel);
const devToolsMiddleware = vi.fn(defaultDevToolsMiddleware);
const provider = createKtxLlmProvider(
{
backend: 'anthropic',
anthropic: { apiKey: 'test-anthropic-key' }, // pragma: allowlist secret
modelSlots: { default: 'claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{
createAnthropic: vi.fn(() => vi.fn(() => anthropicModel)),
devtoolsEnabled: false,
wrapLanguageModel,
devToolsMiddleware,
} satisfies KtxLlmProviderFactoryDeps,
);
expect(provider.getModel('default')).toBe(anthropicModel);
expect(wrapLanguageModel).not.toHaveBeenCalled();
expect(devToolsMiddleware).not.toHaveBeenCalled();
});
it('wraps language models when KTX_AI_DEVTOOLS_ENABLED is true', () => {
const originalEnv = process.env.KTX_AI_DEVTOOLS_ENABLED;
process.env.KTX_AI_DEVTOOLS_ENABLED = 'true';
try {
const gatewayModel = languageModel('anthropic/claude-sonnet-4-6', 'gateway');
const wrappedModel = languageModel('anthropic/claude-sonnet-4-6', 'gateway-devtools');
const wrapLanguageModel = wrapWith(wrappedModel);
const provider = createKtxLlmProvider(
{
backend: 'gateway',
gateway: { baseURL: 'https://gateway.test/v1' },
modelSlots: { default: 'anthropic/claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{
createGateway: vi.fn(() => vi.fn(() => gatewayModel)),
wrapLanguageModel,
devToolsMiddleware: vi.fn(devtoolsMiddleware),
} satisfies KtxLlmProviderFactoryDeps,
);
expect(provider.getModel('default')).toBe(wrappedModel);
expect(wrapLanguageModel).toHaveBeenCalledTimes(1);
} finally {
if (originalEnv === undefined) {
delete process.env.KTX_AI_DEVTOOLS_ENABLED;
} else {
process.env.KTX_AI_DEVTOOLS_ENABLED = originalEnv;
}
}
});
it('does not wrap language models in production even when enabled', () => {
const originalNodeEnv = process.env.NODE_ENV;
process.env.NODE_ENV = 'production';
try {
const anthropicModel = languageModel('claude-sonnet-4-6', 'anthropic');
const wrapLanguageModel = vi.fn(defaultWrapLanguageModel);
const devToolsMiddleware = vi.fn(defaultDevToolsMiddleware);
const provider = createKtxLlmProvider(
{
backend: 'anthropic',
anthropic: { apiKey: 'test-anthropic-key' }, // pragma: allowlist secret
modelSlots: { default: 'claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{
createAnthropic: vi.fn(() => vi.fn(() => anthropicModel)),
devtoolsEnabled: true,
wrapLanguageModel,
devToolsMiddleware,
} satisfies KtxLlmProviderFactoryDeps,
);
expect(provider.getModel('default')).toBe(anthropicModel);
expect(wrapLanguageModel).not.toHaveBeenCalled();
expect(devToolsMiddleware).not.toHaveBeenCalled();
} finally {
if (originalNodeEnv === undefined) {
delete process.env.NODE_ENV;
} else {
process.env.NODE_ENV = originalNodeEnv;
}
}
});
it('uses direct Anthropic with both beta headers', () => {
const anthropicModel = languageModel('claude-sonnet-4-6', 'anthropic');
const anthropic = vi.fn(() => anthropicModel);
const createAnthropic = vi.fn(() => anthropic);
const provider = createKtxLlmProvider(
{
backend: 'anthropic',
anthropic: { apiKey: 'test-anthropic-key', baseURL: 'https://anthropic.test' }, // pragma: allowlist secret
modelSlots: { default: 'claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{ createAnthropic, devtoolsEnabled: false },
);
expect(provider.getModel('default')).toBe(anthropicModel);
expect(createAnthropic).toHaveBeenCalledWith({
apiKey: 'test-anthropic-key', // pragma: allowlist secret
baseURL: 'https://anthropic.test',
headers: {
'anthropic-beta': 'interleaved-thinking-2025-05-14,extended-cache-ttl-2025-04-11',
},
});
expect(anthropic).toHaveBeenCalledWith('claude-sonnet-4-6');
});
it('uses Vertex Anthropic without the direct-Anthropic beta header', () => {
const vertexModel = languageModel('claude-sonnet-4-6', 'vertex');
const vertex = vi.fn(() => vertexModel);
const createVertexAnthropic = vi.fn(() => vertex);
const provider = createKtxLlmProvider(
{
backend: 'vertex',
vertex: { project: 'ktx-test', location: 'us-east5' },
modelSlots: { default: 'claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{ createVertexAnthropic, devtoolsEnabled: false },
);
expect(provider.getModel('default')).toBe(vertexModel);
expect(createVertexAnthropic).toHaveBeenCalledWith({ project: 'ktx-test', location: 'us-east5' });
expect(vertex).toHaveBeenCalledWith('claude-sonnet-4-6');
});
it('uses Gateway and supports role fallback to default', () => {
const gatewayModel = languageModel('anthropic/claude-sonnet-4-6', 'gateway');
const gateway = vi.fn(() => gatewayModel);
const createGateway = vi.fn(() => gateway);
const provider = createKtxLlmProvider(
{
backend: 'gateway',
gateway: { apiKey: 'gateway-key', baseURL: 'https://gateway.test/v1' }, // pragma: allowlist secret
modelSlots: { default: 'anthropic/claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{ createGateway, devtoolsEnabled: false },
);
expect(provider.getModel('curator')).toBe(gatewayModel);
expect(createGateway).toHaveBeenCalledWith({
apiKey: 'gateway-key', // pragma: allowlist secret
baseURL: 'https://gateway.test/v1',
headers: {
'anthropic-beta': 'interleaved-thinking-2025-05-14,extended-cache-ttl-2025-04-11',
},
});
expect(gateway).toHaveBeenCalledWith('anthropic/claude-sonnet-4-6');
});
it('uses explicit role overrides before default', () => {
const anthropic = vi.fn((modelId: string) => languageModel(modelId, 'anthropic'));
const provider = createKtxLlmProvider(
{
backend: 'anthropic',
anthropic: { apiKey: 'test-anthropic-key' }, // pragma: allowlist secret
modelSlots: {
default: 'claude-sonnet-4-6',
triage: 'claude-haiku-4-5',
repair: 'claude-opus-4-7',
},
promptCaching: { enabled: false },
},
{ createAnthropic: vi.fn(() => anthropic) },
);
expect((provider.getModel('triage') as { modelId: string }).modelId).toBe('claude-haiku-4-5');
expect((provider.getModel('repair') as { modelId: string }).modelId).toBe('claude-opus-4-7');
expect((provider.getModel('reconcile') as { modelId: string }).modelId).toBe('claude-sonnet-4-6');
});
it('emits cache markers only when enabled and the model speaks Anthropic protocol', () => {
const provider = createKtxLlmProvider(
{
backend: 'gateway',
gateway: { baseURL: 'https://gateway.test/v1' },
modelSlots: { default: 'anthropic/claude-sonnet-4-6' },
promptCaching: { enabled: true },
},
{ createGateway: vi.fn(() => vi.fn((modelId: string) => languageModel(modelId, 'gateway'))) },
);
expect(provider.cacheMarker('1h', 'anthropic/claude-sonnet-4-6')).toEqual({
anthropic: { cacheControl: { type: 'ephemeral', ttl: '1h' } },
});
expect(provider.cacheMarker('1h', 'gpt-5')).toBeUndefined();
});
it('returns Anthropic thinking provider options', () => {
const provider = createKtxLlmProvider(
{
backend: 'anthropic',
anthropic: { apiKey: 'test-anthropic-key' }, // pragma: allowlist secret
modelSlots: { default: 'claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{ createAnthropic: vi.fn(() => vi.fn((modelId: string) => languageModel(modelId, 'anthropic'))) },
);
expect(provider.thinkingProviderOptions('default', 12000)).toEqual({
anthropic: {
thinking: { type: 'enabled', budgetTokens: 12000 },
},
});
});
it('defaults prompt caching to enabled with canonical TTLs', () => {
const provider = createKtxLlmProvider(
{
backend: 'gateway',
gateway: { baseURL: 'https://gateway.test/v1' },
modelSlots: { default: 'anthropic/claude-sonnet-4-6' },
},
{ createGateway: vi.fn(() => vi.fn((modelId: string) => languageModel(modelId, 'gateway'))) },
);
expect(provider.promptCachingConfig()).toEqual({
enabled: true,
systemTtl: '1h',
toolsTtl: '1h',
historyTtl: '5m',
cacheSystem: true,
cacheTools: true,
cacheHistory: true,
vertexFallbackTo5m: false,
});
expect(provider.cacheMarker('1h', 'anthropic/claude-sonnet-4-6')).toEqual({
anthropic: { cacheControl: { type: 'ephemeral', ttl: '1h' } },
});
});
it('preserves explicit prompt caching opt-out', () => {
const provider = createKtxLlmProvider(
{
backend: 'anthropic',
anthropic: { apiKey: 'test-anthropic-key' }, // pragma: allowlist secret
modelSlots: { default: 'claude-sonnet-4-6' },
promptCaching: { enabled: false },
},
{ createAnthropic: vi.fn(() => vi.fn((modelId: string) => languageModel(modelId, 'anthropic'))) },
);
expect(provider.promptCachingConfig().enabled).toBe(false);
expect(provider.cacheMarker('1h', 'claude-sonnet-4-6')).toBeUndefined();
});
it('throws instead of falling through when an unsupported LLM backend is passed to the AI SDK provider factory', () => {
expect(() =>
createKtxLlmProvider({
backend: 'claude-code',
modelSlots: { default: 'sonnet' },
promptCaching: { enabled: false },
}),
).toThrow('claude-code is not an AI SDK LanguageModel backend');
});
it('rejects codex as an AI SDK LanguageModel backend', () => {
expect(() =>
createKtxLlmProvider({
backend: 'codex',
modelSlots: { default: 'gpt-5.3-codex' },
}),
).toThrow('codex is not an AI SDK LanguageModel backend');
});
});