Compare commits

..

2 Commits

Author SHA1 Message Date
Jarvis
0ef682e053 feat(gateway): honor MOSAIC_GATEWAY_SKIP_NPM_INSTALL to skip registry reinstall
Some checks failed
ci/woodpecker/push/ci Pipeline failed
ci/woodpecker/pr/ci Pipeline was successful
Lets dev/offline installs keep an already-present global @mosaicstack/gateway
(e.g. a build-from-source `install.sh --dev`) instead of overwriting it with
the registry @latest build during `mosaic wizard` / `mosaic gateway install`.

Refs #675

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RMoEx7hfdFGjUiCHuN1RRi
2026-06-24 16:46:03 -05:00
Jarvis
d604d31676 fix(installer): require provider in quick-start (#1) + local-tier gateway Redis gate (#2)
Some checks failed
ci/woodpecker/push/ci Pipeline was successful
ci/woodpecker/pr/ci Pipeline failed
Bug #1 — quick-start wizard let users skip the LLM provider/API key yet
reported "Mosaic is ready". provider-setup now requires an interactive
provider select + a validated key; quick-start guards the headless path;
finalize won't print "Mosaic is ready" without a configured provider;
removed references to the non-existent `mosaic configure` command.

Bug #2 — the "local" storage tier still opened ioredis/BullMQ
connections at bootstrap (ECONNREFUSED, gateway never healthy). Every
Redis consumer (queue.service, gc.module, session-gc.service,
commands.module, command-executor.service, system-override.service,
cron.service, admin-health.controller) is now tier-aware and degrades
gracefully on local tier. Standalone/Federated unaffected. Also fixed a
pre-existing SystemOverrideService ioredis handle leak (added shutdown hook).

Refs #675

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RMoEx7hfdFGjUiCHuN1RRi
2026-06-24 16:35:23 -05:00
27 changed files with 376 additions and 920 deletions

View File

@@ -1,9 +1,11 @@
import { Controller, Get, Inject, UseGuards } from '@nestjs/common';
import { Controller, Get, Inject, Optional, UseGuards } from '@nestjs/common';
import { sql, type Db } from '@mosaicstack/db';
import { createQueue } from '@mosaicstack/queue';
import type { MosaicConfig } from '@mosaicstack/config';
import { DB } from '../database/database.module.js';
import { AgentService } from '../agent/agent.service.js';
import { ProviderService } from '../agent/provider.service.js';
import { MOSAIC_CONFIG } from '../config/config.module.js';
import { AdminGuard } from './admin.guard.js';
import type { HealthStatusDto, ServiceStatusDto } from './admin.dto.js';
@@ -14,6 +16,9 @@ export class AdminHealthController {
@Inject(DB) private readonly db: Db,
@Inject(AgentService) private readonly agentService: AgentService,
@Inject(ProviderService) private readonly providerService: ProviderService,
@Optional()
@Inject(MOSAIC_CONFIG)
private readonly mosaicConfig: MosaicConfig | null,
) {}
@Get()
@@ -55,6 +60,14 @@ export class AdminHealthController {
}
private async checkCache(): Promise<ServiceStatusDto> {
// On Local tier there is no Redis. The cache is intentionally absent, which
// is a healthy state for this tier — report 'ok' rather than opening a new
// ioredis connection on every admin health check (which would spam
// ECONNREFUSED and create/destroy a connection per request). latencyMs 0
// signals "no cache backend to measure" for this tier.
if (this.mosaicConfig?.queue?.type === 'local') {
return { status: 'ok', latencyMs: 0 };
}
const start = Date.now();
const handle = createQueue();
try {

View File

@@ -21,7 +21,10 @@ export class CommandExecutorService {
@Inject(AgentService) private readonly agentService: AgentService,
@Inject(SystemOverrideService) private readonly systemOverride: SystemOverrideService,
@Inject(SessionGCService) private readonly sessionGC: SessionGCService,
@Inject(COMMANDS_REDIS) private readonly redis: QueueHandle['redis'],
// On Local tier COMMANDS_REDIS is null — provider login caching is skipped.
@Optional()
@Inject(COMMANDS_REDIS)
private readonly redis: QueueHandle['redis'] | null,
@Inject(BRAIN) private readonly brain: Brain,
@Optional()
@Inject(forwardRef(() => ReloadService))
@@ -403,14 +406,16 @@ export class CommandExecutorService {
};
}
const pollToken = crypto.randomUUID();
const key = `mosaic:auth:poll:${pollToken}`;
// Store pending state in Valkey (TTL 5 minutes)
await this.redis.set(
key,
JSON.stringify({ status: 'pending', provider: providerName, userId }),
'EX',
300,
);
const pollKey = `mosaic:auth:poll:${pollToken}`;
if (this.redis) {
// Store pending state in Valkey (TTL 5 minutes)
await this.redis.set(
pollKey,
JSON.stringify({ status: 'pending', provider: providerName, userId }),
'EX',
300,
);
}
// In production this would construct an OAuth URL
const loginUrl = `${process.env['MOSAIC_BASE_URL'] ?? 'http://localhost:3000'}/auth/provider/${providerName}?token=${pollToken}`;
return {

View File

@@ -1,5 +1,7 @@
import { forwardRef, Inject, Module, type OnApplicationShutdown } from '@nestjs/common';
import { forwardRef, Inject, Module, Optional, type OnApplicationShutdown } from '@nestjs/common';
import { createQueue, type QueueHandle } from '@mosaicstack/queue';
import type { MosaicConfig } from '@mosaicstack/config';
import { MOSAIC_CONFIG } from '../config/config.module.js';
import { ChatModule } from '../chat/chat.module.js';
import { GCModule } from '../gc/gc.module.js';
import { ReloadModule } from '../reload/reload.module.js';
@@ -14,13 +16,17 @@ const COMMANDS_QUEUE_HANDLE = 'COMMANDS_QUEUE_HANDLE';
providers: [
{
provide: COMMANDS_QUEUE_HANDLE,
useFactory: (): QueueHandle => {
useFactory: (config: MosaicConfig | null): QueueHandle | null => {
// On Local tier there is no Redis — skip the ioredis connection.
// CommandExecutorService falls back to no-cache for /provider login on local.
if (config?.queue?.type === 'local') return null;
return createQueue();
},
inject: [MOSAIC_CONFIG],
},
{
provide: COMMANDS_REDIS,
useFactory: (handle: QueueHandle) => handle.redis,
useFactory: (handle: QueueHandle | null) => handle?.redis ?? null,
inject: [COMMANDS_QUEUE_HANDLE],
},
CommandRegistryService,
@@ -29,9 +35,13 @@ const COMMANDS_QUEUE_HANDLE = 'COMMANDS_QUEUE_HANDLE';
exports: [CommandRegistryService, CommandExecutorService],
})
export class CommandsModule implements OnApplicationShutdown {
constructor(@Inject(COMMANDS_QUEUE_HANDLE) private readonly handle: QueueHandle) {}
constructor(
@Optional()
@Inject(COMMANDS_QUEUE_HANDLE)
private readonly handle: QueueHandle | null,
) {}
async onApplicationShutdown(): Promise<void> {
await this.handle.close().catch(() => {});
await this.handle?.close().catch(() => {});
}
}

View File

@@ -1,255 +0,0 @@
import 'reflect-metadata';
import { describe, expect, it, vi } from 'vitest';
import type { Db } from '@mosaicstack/db';
import type { FederationListResponse } from '@mosaicstack/types';
import {
FederationClientError,
type FederationClientService,
} from '../federation-client.service.js';
import { type QuerySourceError, QuerySourceService } from '../query-source.service.js';
interface TestRow {
id: string;
title: string;
}
interface PeerRow {
id: string;
commonName: string;
endpointUrl: string | null;
clientKeyPem: string | null;
state: 'active' | 'pending' | 'suspended' | 'revoked';
}
const LOCAL_ROWS: TestRow[] = [
{ id: 'local-1', title: 'Local One' },
{ id: 'local-2', title: 'Local Two' },
];
const PEER_A: PeerRow = {
id: 'aaaaaaaa-aaaa-aaaa-aaaa-aaaaaaaaaaaa',
commonName: 'peer-a',
endpointUrl: 'https://peer-a.example.com',
clientKeyPem: 'sealed-key-a',
state: 'active',
};
const PEER_B: PeerRow = {
id: 'bbbbbbbb-bbbb-bbbb-bbbb-bbbbbbbbbbbb',
commonName: 'peer-b',
endpointUrl: 'https://peer-b.example.com',
clientKeyPem: 'sealed-key-b',
state: 'active',
};
const PEER_LOCALHOST: PeerRow = {
id: 'cccccccc-cccc-cccc-cccc-cccccccccccc',
commonName: 'peer-localhost',
endpointUrl: 'https://localhost:3001',
clientKeyPem: 'sealed-key-c',
state: 'active',
};
function makeDb(activePeers: PeerRow[]): Db {
const orderBy = vi.fn().mockResolvedValue(activePeers);
const where = vi.fn().mockReturnValue({ orderBy });
const from = vi.fn().mockReturnValue({ where });
const select = vi.fn().mockReturnValue({ from });
return {
select,
insert: vi.fn(),
update: vi.fn(),
delete: vi.fn(),
transaction: vi.fn(),
} as unknown as Db;
}
function makeFederationClient(
list: (
peerId: string,
resource: string,
request: Record<string, unknown>,
) => Promise<FederationListResponse<TestRow>>,
): FederationClientService {
return {
list: list as unknown as FederationClientService['list'],
} as FederationClientService;
}
function makeLocalResponse(rows: TestRow[] = LOCAL_ROWS): Promise<FederationListResponse<TestRow>> {
return Promise.resolve({ items: rows });
}
describe('QuerySourceService', () => {
it('routes source="local" to the local executor and tags rows as local', async () => {
const list = vi.fn(async (): Promise<FederationListResponse<TestRow>> => ({ items: [] }));
const service = new QuerySourceService(makeDb([PEER_A]), makeFederationClient(list));
const result = await service.list<TestRow>({
source: 'local',
resource: 'tasks',
request: { cursor: 'ignored-for-local-test' },
local: () => makeLocalResponse(),
});
expect(result).toEqual({
items: [
{ id: 'local-1', title: 'Local One', _source: 'local' },
{ id: 'local-2', title: 'Local Two', _source: 'local' },
],
});
expect(list).not.toHaveBeenCalled();
});
it('routes source="federated:<host>" to the matching active peer and tags rows with peer commonName', async () => {
const list = vi.fn(
async (): Promise<FederationListResponse<TestRow>> => ({
items: [{ id: 'remote-1', title: 'Remote One' }],
}),
);
const service = new QuerySourceService(makeDb([PEER_A, PEER_B]), makeFederationClient(list));
const result = await service.list<TestRow>({
source: 'federated:peer-b.example.com',
resource: 'tasks',
request: { status: 'open' },
local: () => makeLocalResponse(),
});
expect(result).toEqual({
items: [{ id: 'remote-1', title: 'Remote One', _source: 'peer-b' }],
});
expect(list).toHaveBeenCalledWith(PEER_B.id, 'tasks', { status: 'open' });
});
it('matches federated hosts by endpoint host including non-default port', async () => {
const list = vi.fn(
async (): Promise<FederationListResponse<TestRow>> => ({
items: [{ id: 'remote-port', title: 'Remote Port' }],
}),
);
const service = new QuerySourceService(makeDb([PEER_LOCALHOST]), makeFederationClient(list));
const result = await service.list<TestRow>({
source: 'federated:localhost:3001',
resource: 'tasks',
request: {},
local: () => makeLocalResponse(),
});
expect(result).toEqual({
items: [{ id: 'remote-port', title: 'Remote Port', _source: 'peer-localhost' }],
});
expect(list).toHaveBeenCalledWith(PEER_LOCALHOST.id, 'tasks', {});
});
it('fans out source="all" to local plus every active outbound peer in parallel and merges tagged rows', async () => {
const callOrder: string[] = [];
const list = vi.fn(async (peerId: string): Promise<FederationListResponse<TestRow>> => {
callOrder.push(`remote-start:${peerId}`);
await Promise.resolve();
return {
items: [{ id: `remote-${peerId.slice(0, 1)}`, title: `Remote ${peerId.slice(0, 1)}` }],
};
});
const service = new QuerySourceService(makeDb([PEER_A, PEER_B]), makeFederationClient(list));
const result = await service.list<TestRow>({
source: 'all',
resource: 'tasks',
request: { limit: 25 },
local: async () => {
callOrder.push('local-start');
await Promise.resolve();
return { items: [{ id: 'local-1', title: 'Local One' }] };
},
});
expect(result).toEqual({
items: [
{ id: 'local-1', title: 'Local One', _source: 'local' },
{ id: 'remote-a', title: 'Remote a', _source: 'peer-a' },
{ id: 'remote-b', title: 'Remote b', _source: 'peer-b' },
],
});
expect(list).toHaveBeenCalledTimes(2);
expect(callOrder).toEqual([
'local-start',
`remote-start:${PEER_A.id}`,
`remote-start:${PEER_B.id}`,
]);
});
it('marks source="all" as partial and truncated when any subquery returns a cursor', async () => {
const list = vi.fn(
async (): Promise<FederationListResponse<TestRow>> => ({
items: [{ id: 'remote-a', title: 'Remote A' }],
nextCursor: 'remote-next',
}),
);
const service = new QuerySourceService(makeDb([PEER_A]), makeFederationClient(list));
const result = await service.list<TestRow>({
source: 'all',
resource: 'tasks',
request: {},
local: () => makeLocalResponse([{ id: 'local-1', title: 'Local One' }]),
});
expect(result).toEqual({
items: [
{ id: 'local-1', title: 'Local One', _source: 'local' },
{ id: 'remote-a', title: 'Remote A', _source: 'peer-a' },
],
_partial: true,
_truncated: true,
});
});
it('returns _partial=true for source="all" when one peer fails without dropping successful sources', async () => {
const list = vi.fn(async (peerId: string): Promise<FederationListResponse<TestRow>> => {
if (peerId === PEER_B.id) {
throw new FederationClientError({
code: 'NETWORK',
message: 'peer unavailable',
peerId,
});
}
return { items: [{ id: 'remote-a', title: 'Remote A' }] };
});
const service = new QuerySourceService(makeDb([PEER_A, PEER_B]), makeFederationClient(list));
const result = await service.list<TestRow>({
source: 'all',
resource: 'tasks',
request: {},
local: () => makeLocalResponse([{ id: 'local-1', title: 'Local One' }]),
});
expect(result).toEqual({
items: [
{ id: 'local-1', title: 'Local One', _source: 'local' },
{ id: 'remote-a', title: 'Remote A', _source: 'peer-a' },
],
_partial: true,
});
});
it('throws QuerySourceError when a federated host does not match an active outbound peer', async () => {
const list = vi.fn(async (): Promise<FederationListResponse<TestRow>> => ({ items: [] }));
const service = new QuerySourceService(makeDb([PEER_A]), makeFederationClient(list));
await expect(
service.list<TestRow>({
source: 'federated:missing.example.com',
resource: 'tasks',
request: {},
local: () => makeLocalResponse(),
}),
).rejects.toMatchObject({
name: 'QuerySourceError',
code: 'PEER_NOT_FOUND',
} satisfies Partial<QuerySourceError>);
});
});

View File

@@ -11,13 +11,3 @@ export {
type FederationClientErrorCode,
type FederationClientErrorOptions,
} from './federation-client.service.js';
export {
QuerySourceService,
QuerySourceError,
type QuerySource,
type QuerySourceErrorCode,
type QuerySourceErrorOptions,
type QuerySourceListOptions,
type QuerySourceListResponse,
type LocalListExecutor,
} from './query-source.service.js';

View File

@@ -1,261 +0,0 @@
/**
* QuerySourceService — gateway query source router (FED-M3-09).
*
* Accepts the federation query-layer `source` selector and routes list-style
* reads to local storage, one federated peer, or all active outbound peers.
*
* `source: "all"` is intentionally tolerant of per-peer failures: local data
* and successful peer responses are returned, and the envelope is marked
* `_partial: true`. Local failures still reject because there is no safe local
* fallback and the gateway's own storage is expected to be authoritative.
*/
import { Inject, Injectable, Logger } from '@nestjs/common';
import { and, eq, federationPeers, isNotNull, type Db } from '@mosaicstack/db';
import {
SOURCE_LOCAL,
tagWithSource,
type FederationListResponse,
type SourceTag,
} from '@mosaicstack/types';
import { DB } from '../../database/database.module.js';
import { FederationClientService } from './federation-client.service.js';
export type QuerySource = 'local' | 'all' | `federated:${string}`;
export type QuerySourceErrorCode = 'INVALID_SOURCE' | 'PEER_NOT_FOUND';
export interface QuerySourceErrorOptions {
code: QuerySourceErrorCode;
message: string;
source: string;
}
export class QuerySourceError extends Error {
readonly code: QuerySourceErrorCode;
readonly source: string;
constructor(opts: QuerySourceErrorOptions) {
super(opts.message);
this.name = 'QuerySourceError';
this.code = opts.code;
this.source = opts.source;
}
}
export type LocalListExecutor<T extends object> = () => Promise<FederationListResponse<T> | T[]>;
export interface QuerySourceListOptions<T extends object> {
source: QuerySource;
resource: string;
request?: Record<string, unknown>;
local: LocalListExecutor<T>;
}
export type QuerySourceListResponse<T extends object> = FederationListResponse<T & SourceTag>;
interface OutboundPeer {
id: string;
commonName: string;
endpointUrl: string;
}
interface TaggedList<T extends object> {
items: Array<T & SourceTag>;
partial: boolean;
truncated: boolean;
nextCursor?: string;
}
@Injectable()
export class QuerySourceService {
private readonly logger = new Logger(QuerySourceService.name);
constructor(
@Inject(DB) private readonly db: Db,
@Inject(FederationClientService) private readonly federationClient: FederationClientService,
) {}
async list<T extends object>(
options: QuerySourceListOptions<T>,
): Promise<QuerySourceListResponse<T>> {
const request = options.request ?? {};
if (options.source === 'local') {
const local = await this.runLocal(options.local);
return this.toResponse(this.tagList(local, SOURCE_LOCAL));
}
if (options.source === 'all') {
return this.listAll(options.resource, request, options.local);
}
if (options.source.startsWith('federated:')) {
const host = options.source.slice('federated:'.length).trim();
if (!host) {
throw new QuerySourceError({
code: 'INVALID_SOURCE',
message: 'Federated source must include a host after federated:',
source: options.source,
});
}
const peer = await this.findPeerByHost(host, options.source);
const remote = await this.federationClient.list<T>(peer.id, options.resource, request);
return this.toResponse(this.tagList(remote, peer.commonName));
}
throw new QuerySourceError({
code: 'INVALID_SOURCE',
message: `Unsupported query source: ${options.source}`,
source: options.source,
});
}
private async listAll<T extends object>(
resource: string,
request: Record<string, unknown>,
local: LocalListExecutor<T>,
): Promise<QuerySourceListResponse<T>> {
const peers = await this.listActiveOutboundPeers();
const localPromise = this.runLocal(local).then((response) =>
this.tagList(response, SOURCE_LOCAL),
);
const remotePromises = peers.map(async (peer: OutboundPeer): Promise<TaggedList<T> | null> => {
try {
const response = await this.federationClient.list<T>(peer.id, resource, request);
return this.tagList(response, peer.commonName);
} catch (error: unknown) {
this.logger.warn(
`Federated query to peer ${peer.commonName} (${peer.id}) failed; returning partial all-source response: ${
error instanceof Error ? error.message : String(error)
}`,
);
return null;
}
});
const [localResult, ...remoteResults] = await Promise.all([localPromise, ...remotePromises]);
const successfulRemoteResults = remoteResults.filter(
(result: TaggedList<T> | null): result is TaggedList<T> => result !== null,
);
const allResults = [localResult, ...successfulRemoteResults];
const peerFailure = successfulRemoteResults.length !== peers.length;
return this.mergeTaggedLists(allResults, peerFailure);
}
private async runLocal<T extends object>(
local: LocalListExecutor<T>,
): Promise<FederationListResponse<T>> {
const response = await local();
if (Array.isArray(response)) {
return { items: response };
}
return response;
}
private tagList<T extends object>(
response: FederationListResponse<T>,
source: string,
): TaggedList<T> {
return {
items: tagWithSource(response.items, source),
partial: response._partial === true,
truncated: response._truncated === true || response.nextCursor !== undefined,
nextCursor: response.nextCursor,
};
}
private mergeTaggedLists<T extends object>(
lists: Array<TaggedList<T>>,
peerFailure: boolean,
): QuerySourceListResponse<T> {
const items = lists.flatMap((list: TaggedList<T>) => list.items);
const partial =
peerFailure ||
lists.some((list: TaggedList<T>) => list.partial || list.nextCursor !== undefined);
const truncated = lists.some((list: TaggedList<T>) => list.truncated);
const response: QuerySourceListResponse<T> = { items };
if (partial) {
response._partial = true;
}
if (truncated) {
response._truncated = true;
}
return response;
}
private toResponse<T extends object>(tagged: TaggedList<T>): QuerySourceListResponse<T> {
const response: QuerySourceListResponse<T> = {
items: tagged.items,
};
if (tagged.nextCursor !== undefined) {
response.nextCursor = tagged.nextCursor;
}
if (tagged.partial) {
response._partial = true;
}
if (tagged.truncated) {
response._truncated = true;
}
return response;
}
private async findPeerByHost(sourceHost: string, source: string): Promise<OutboundPeer> {
const host = normalizeHost(sourceHost);
const peers = await this.listActiveOutboundPeers();
const peer = peers.find((candidate: OutboundPeer) => {
const commonName = normalizeHost(candidate.commonName);
const endpointHosts = endpointHostKeys(candidate.endpointUrl).map((endpointHost: string) =>
normalizeHost(endpointHost),
);
return commonName === host || endpointHosts.includes(host);
});
if (!peer) {
throw new QuerySourceError({
code: 'PEER_NOT_FOUND',
message: `No active outbound federation peer matches source ${source}`,
source,
});
}
return peer;
}
private async listActiveOutboundPeers(): Promise<OutboundPeer[]> {
const rows = await this.db
.select({
id: federationPeers.id,
commonName: federationPeers.commonName,
endpointUrl: federationPeers.endpointUrl,
})
.from(federationPeers)
.where(
and(
eq(federationPeers.state, 'active'),
isNotNull(federationPeers.endpointUrl),
isNotNull(federationPeers.clientKeyPem),
),
)
.orderBy(federationPeers.commonName);
return rows.filter((row): row is OutboundPeer => typeof row.endpointUrl === 'string');
}
}
function normalizeHost(host: string): string {
return host.trim().toLowerCase();
}
function endpointHostKeys(endpointUrl: string): string[] {
try {
const url = new URL(endpointUrl);
return Array.from(new Set([url.host, url.hostname].filter((host: string) => host.length > 0)));
} catch {
return [];
}
}

View File

@@ -4,20 +4,18 @@ import { CaService } from './ca.service.js';
import { EnrollmentController } from './enrollment.controller.js';
import { EnrollmentService } from './enrollment.service.js';
import { FederationController } from './federation.controller.js';
import { CapabilitiesController } from './server/verbs/capabilities.controller.js';
import { GrantsService } from './grants.service.js';
import { FederationClientService, QuerySourceService } from './client/index.js';
import { FederationClientService } from './client/index.js';
import { FederationAuthGuard } from './server/index.js';
@Module({
controllers: [EnrollmentController, FederationController, CapabilitiesController],
controllers: [EnrollmentController, FederationController],
providers: [
AdminGuard,
CaService,
EnrollmentService,
GrantsService,
FederationClientService,
QuerySourceService,
FederationAuthGuard,
],
exports: [
@@ -25,7 +23,6 @@ import { FederationAuthGuard } from './server/index.js';
EnrollmentService,
GrantsService,
FederationClientService,
QuerySourceService,
FederationAuthGuard,
],
})

View File

@@ -1,88 +0,0 @@
import 'reflect-metadata';
import { RequestMethod } from '@nestjs/common';
import { describe, expect, it } from 'vitest';
import type { FastifyRequest } from 'fastify';
import { FederationCapabilitiesResponseSchema, FEDERATION_VERBS } from '@mosaicstack/types';
import { FederationScopeError } from '../../../scope-schema.js';
import { FederationAuthGuard } from '../../federation-auth.guard.js';
import { CapabilitiesController } from '../capabilities.controller.js';
const VALID_SCOPE = {
resources: ['tasks', 'notes'],
excluded_resources: ['credentials'],
max_rows_per_query: 250,
} as const;
const DEFAULTED_SCOPE = {
resources: ['memory'],
max_rows_per_query: 10,
} as const;
function makeRequest(scope: Record<string, unknown>): FastifyRequest {
return {
federationContext: {
grantId: 'grant-1',
peerId: 'peer-1',
subjectUserId: 'user-1',
scope,
},
} as FastifyRequest;
}
describe('CapabilitiesController', () => {
it('declares GET /api/federation/v1/capabilities', () => {
expect(Reflect.getMetadata('path', CapabilitiesController)).toBe(
'api/federation/v1/capabilities',
);
expect(Reflect.getMetadata('path', CapabilitiesController.prototype.getCapabilities)).toBe('/');
expect(Reflect.getMetadata('method', CapabilitiesController.prototype.getCapabilities)).toBe(
RequestMethod.GET,
);
});
it('is protected only by FederationAuthGuard', () => {
const guards = Reflect.getMetadata('__guards__', CapabilitiesController) as unknown[];
expect(guards).toEqual([FederationAuthGuard]);
});
it('returns resources, excluded resources, max rows, and M3 supported verbs from the active grant scope', () => {
const controller = new CapabilitiesController();
const response = controller.getCapabilities(makeRequest(VALID_SCOPE));
expect(response).toEqual({
resources: ['tasks', 'notes'],
excluded_resources: ['credentials'],
max_rows_per_query: 250,
supported_verbs: [...FEDERATION_VERBS],
});
expect(FederationCapabilitiesResponseSchema.safeParse(response).success).toBe(true);
});
it('applies scope defaults without RBAC or resource filtering', () => {
const controller = new CapabilitiesController();
const response = controller.getCapabilities(makeRequest(DEFAULTED_SCOPE));
expect(response).toEqual({
resources: ['memory'],
excluded_resources: [],
max_rows_per_query: 10,
supported_verbs: ['list', 'get', 'capabilities'],
});
});
it('rejects invalid scope state instead of returning an invalid capabilities contract', () => {
const controller = new CapabilitiesController();
expect(() =>
controller.getCapabilities(
makeRequest({
resources: [],
max_rows_per_query: 0,
}),
),
).toThrow(FederationScopeError);
});
});

View File

@@ -1,38 +0,0 @@
/**
* Federation capabilities verb (FED-M3-07).
*
* Returns the read-only capability envelope for the active grant attached by
* FederationAuthGuard. This endpoint intentionally does not invoke native RBAC
* or ScopeService: an active grant is enough to ask what the grant allows.
*/
import { Controller, Get, Req, UseGuards } from '@nestjs/common';
import type { FastifyRequest } from 'fastify';
import {
FEDERATION_VERBS,
type FederationCapabilitiesResponse,
type FederationVerb,
} from '@mosaicstack/types';
import { parseFederationScope } from '../../scope-schema.js';
import { FederationAuthGuard } from '../federation-auth.guard.js';
import '../federation-context.js';
@Controller('api/federation/v1/capabilities')
@UseGuards(FederationAuthGuard)
export class CapabilitiesController {
@Get()
getCapabilities(@Req() request: FastifyRequest): FederationCapabilitiesResponse {
if (!request.federationContext) {
throw new Error('Federation context missing after auth guard');
}
const scope = parseFederationScope(request.federationContext.scope);
return {
resources: [...scope.resources],
excluded_resources: [...scope.excluded_resources],
max_rows_per_query: scope.max_rows_per_query,
supported_verbs: [...FEDERATION_VERBS] satisfies FederationVerb[],
};
}
}

View File

@@ -1,5 +1,7 @@
import { Module, type OnApplicationShutdown, Inject } from '@nestjs/common';
import { Module, type OnApplicationShutdown, Inject, Optional } from '@nestjs/common';
import { createQueue, type QueueHandle } from '@mosaicstack/queue';
import type { MosaicConfig } from '@mosaicstack/config';
import { MOSAIC_CONFIG } from '../config/config.module.js';
import { SessionGCService } from './session-gc.service.js';
import { REDIS } from './gc.tokens.js';
@@ -9,13 +11,17 @@ const GC_QUEUE_HANDLE = 'GC_QUEUE_HANDLE';
providers: [
{
provide: GC_QUEUE_HANDLE,
useFactory: (): QueueHandle => {
useFactory: (config: MosaicConfig | null): QueueHandle | null => {
// On Local tier there is no Redis — skip the ioredis connection entirely.
// The Valkey GC sweep is a no-op on Local (no session keys stored there).
if (config?.queue?.type === 'local') return null;
return createQueue();
},
inject: [MOSAIC_CONFIG],
},
{
provide: REDIS,
useFactory: (handle: QueueHandle) => handle.redis,
useFactory: (handle: QueueHandle | null) => handle?.redis ?? null,
inject: [GC_QUEUE_HANDLE],
},
SessionGCService,
@@ -23,9 +29,13 @@ const GC_QUEUE_HANDLE = 'GC_QUEUE_HANDLE';
exports: [SessionGCService],
})
export class GCModule implements OnApplicationShutdown {
constructor(@Inject(GC_QUEUE_HANDLE) private readonly handle: QueueHandle) {}
constructor(
@Optional()
@Inject(GC_QUEUE_HANDLE)
private readonly handle: QueueHandle | null,
) {}
async onApplicationShutdown(): Promise<void> {
await this.handle.close().catch(() => {});
await this.handle?.close().catch(() => {});
}
}

View File

@@ -1,4 +1,4 @@
import { Inject, Injectable, Logger, type OnModuleInit } from '@nestjs/common';
import { Inject, Injectable, Logger, Optional, type OnModuleInit } from '@nestjs/common';
import type { QueueHandle } from '@mosaicstack/queue';
import type { LogService } from '@mosaicstack/log';
import { LOG_SERVICE } from '../log/log.tokens.js';
@@ -32,11 +32,21 @@ export class SessionGCService implements OnModuleInit {
private readonly logger = new Logger(SessionGCService.name);
constructor(
@Inject(REDIS) private readonly redis: QueueHandle['redis'],
// On Local tier there is no Redis — the GC module provides null for this token.
// NOTE: if a future feature stores Redis-backed state on Local tier, this guard
// would silently skip GC for those keys. Revisit when that happens.
@Optional()
@Inject(REDIS)
private readonly redis: QueueHandle['redis'] | null,
@Inject(LOG_SERVICE) private readonly logService: LogService,
) {}
onModuleInit(): void {
if (!this.redis) {
// Local tier: no Valkey — skip cold-start GC entirely (correct no-op).
this.logger.log('SessionGCService: Valkey GC skipped on local tier (no Redis configured)');
return;
}
// Fire-and-forget: run full GC asynchronously so it does not block the
// NestJS bootstrap chain. Cold-start GC typically takes 100500 ms
// depending on Valkey key count; deferring it removes that latency from
@@ -60,8 +70,10 @@ export class SessionGCService implements OnModuleInit {
* Scan Valkey for all keys matching a pattern using SCAN (non-blocking).
* KEYS is avoided because it blocks the Valkey event loop for the full scan
* duration, which can cause latency spikes under production key volumes.
* Returns empty array when Redis is not available (Local tier).
*/
private async scanKeys(pattern: string): Promise<string[]> {
if (!this.redis) return [];
const collected: string[] = [];
let cursor = '0';
do {
@@ -78,12 +90,14 @@ export class SessionGCService implements OnModuleInit {
async collect(sessionId: string): Promise<GCResult> {
const result: GCResult = { sessionId, cleaned: {} };
// 1. Valkey: delete all session-scoped keys
const pattern = `mosaic:session:${sessionId}:*`;
const valkeyKeys = await this.scanKeys(pattern);
if (valkeyKeys.length > 0) {
await this.redis.del(...valkeyKeys);
result.cleaned.valkeyKeys = valkeyKeys.length;
// 1. Valkey: delete all session-scoped keys (skipped on Local tier)
if (this.redis) {
const pattern = `mosaic:session:${sessionId}:*`;
const valkeyKeys = await this.scanKeys(pattern);
if (valkeyKeys.length > 0) {
await this.redis.del(...valkeyKeys);
result.cleaned.valkeyKeys = valkeyKeys.length;
}
}
// 2. PG: demote hot-tier agent_logs for this session to warm
@@ -106,6 +120,7 @@ export class SessionGCService implements OnModuleInit {
const cleaned: GCResult[] = [];
// 1. Find all session-scoped Valkey keys (non-blocking SCAN)
// Returns empty on Local tier — no Valkey session keys exist there.
const allSessionKeys = await this.scanKeys('mosaic:session:*');
// Extract unique session IDs from keys
@@ -136,11 +151,15 @@ export class SessionGCService implements OnModuleInit {
*/
async fullCollect(): Promise<FullGCResult> {
const start = Date.now();
let valkeyKeysCount = 0;
// 1. Valkey: delete ALL session-scoped keys (non-blocking SCAN)
const sessionKeys = await this.scanKeys('mosaic:session:*');
if (sessionKeys.length > 0) {
await this.redis.del(...sessionKeys);
if (this.redis) {
// 1. Valkey: delete ALL session-scoped keys (non-blocking SCAN)
const sessionKeys = await this.scanKeys('mosaic:session:*');
if (sessionKeys.length > 0) {
await this.redis.del(...sessionKeys);
}
valkeyKeysCount = sessionKeys.length;
}
// 2. NOTE: channel keys are NOT collected on cold start
@@ -154,7 +173,7 @@ export class SessionGCService implements OnModuleInit {
const jobsPurged = 0;
return {
valkeyKeys: sessionKeys.length,
valkeyKeys: valkeyKeysCount,
logsDemoted,
jobsPurged,
tempFilesRemoved: 0,

View File

@@ -19,7 +19,7 @@ import type { MosaicJobData } from '../queue/queue.service.js';
@Injectable()
export class CronService implements OnModuleInit, OnModuleDestroy {
private readonly logger = new Logger(CronService.name);
private readonly registeredWorkers: Worker<MosaicJobData>[] = [];
private readonly registeredWorkers: Array<Worker<MosaicJobData>> = [];
constructor(
@Inject(SummarizationService) private readonly summarization: SummarizationService,
@@ -28,6 +28,16 @@ export class CronService implements OnModuleInit, OnModuleDestroy {
) {}
async onModuleInit(): Promise<void> {
// On Local tier BullMQ is disabled — skip all job scheduling.
// NOTE: this means summarization, tier management, and Valkey GC jobs do not
// run on Local installs. For a single-user local install this is acceptable.
// If periodic background work is needed on Local in the future, add a
// setInterval-based scheduler here.
if (!this.queueService.isEnabled()) {
this.logger.log('CronService: BullMQ disabled on local tier — no jobs will be scheduled');
return;
}
const summarizationSchedule = process.env['SUMMARIZATION_CRON'] ?? '0 */6 * * *'; // every 6 hours
const tierManagementSchedule = process.env['TIER_MANAGEMENT_CRON'] ?? '0 3 * * *'; // daily at 3am
const gcSchedule = process.env['SESSION_GC_CRON'] ?? '0 4 * * *'; // daily at 4am
@@ -42,7 +52,7 @@ export class CronService implements OnModuleInit, OnModuleDestroy {
const summarizationWorker = this.queueService.registerWorker(QUEUE_SUMMARIZATION, async () => {
await this.summarization.runSummarization();
});
this.registeredWorkers.push(summarizationWorker);
if (summarizationWorker) this.registeredWorkers.push(summarizationWorker);
// M6-005: Tier management repeatable job
await this.queueService.addRepeatableJob(
@@ -54,14 +64,14 @@ export class CronService implements OnModuleInit, OnModuleDestroy {
const tierWorker = this.queueService.registerWorker(QUEUE_TIER_MANAGEMENT, async () => {
await this.summarization.runTierManagement();
});
this.registeredWorkers.push(tierWorker);
if (tierWorker) this.registeredWorkers.push(tierWorker);
// M6-004: GC repeatable job
await this.queueService.addRepeatableJob(QUEUE_GC, 'session-gc', {}, gcSchedule);
const gcWorker = this.queueService.registerWorker(QUEUE_GC, async () => {
await this.sessionGC.sweepOrphans();
});
this.registeredWorkers.push(gcWorker);
if (gcWorker) this.registeredWorkers.push(gcWorker);
this.logger.log(
`BullMQ jobs scheduled: summarization="${summarizationSchedule}", tier="${tierManagementSchedule}", gc="${gcSchedule}"`,

View File

@@ -1,5 +1,7 @@
import { Injectable, Logger } from '@nestjs/common';
import { Inject, Injectable, Logger, Optional, type OnApplicationShutdown } from '@nestjs/common';
import { createQueue, type QueueHandle } from '@mosaicstack/queue';
import type { MosaicConfig } from '@mosaicstack/config';
import { MOSAIC_CONFIG } from '../config/config.module.js';
const SESSION_SYSTEM_KEY = (sessionId: string) => `mosaic:session:${sessionId}:system`;
const SESSION_SYSTEM_FRAGMENTS_KEY = (sessionId: string) =>
@@ -11,16 +13,54 @@ interface OverrideFragment {
addedAt: number;
}
@Injectable()
export class SystemOverrideService {
private readonly logger = new Logger(SystemOverrideService.name);
private readonly handle: QueueHandle;
interface LocalOverrideEntry {
condensed: string;
fragments: OverrideFragment[];
}
constructor() {
this.handle = createQueue();
@Injectable()
export class SystemOverrideService implements OnApplicationShutdown {
private readonly logger = new Logger(SystemOverrideService.name);
private readonly handle: QueueHandle | null;
/**
* In-memory fallback used on Local tier (no Redis).
* NOTE: state is ephemeral — lost on restart. For Local single-user installs
* this is acceptable; system overrides are re-applied at the next session.
* This is a deliberate behavior change from the Redis-backed 7-day TTL.
*/
private readonly localStore = new Map<string, LocalOverrideEntry>();
constructor(
@Optional()
@Inject(MOSAIC_CONFIG)
private readonly mosaicConfig: MosaicConfig | null,
) {
if (this.mosaicConfig?.queue?.type === 'local') {
this.handle = null;
} else {
this.handle = createQueue();
}
}
async onApplicationShutdown(): Promise<void> {
// On non-local tiers the constructor opens an ioredis connection; close it
// on graceful shutdown to avoid leaking the handle (local tier is null).
await this.handle?.close().catch(() => {});
}
async set(sessionId: string, override: string): Promise<void> {
if (!this.handle) {
// Local tier: in-memory path
const entry = this.localStore.get(sessionId) ?? { condensed: '', fragments: [] };
entry.fragments.push({ text: override, addedAt: Date.now() });
entry.condensed = await this.condenseOverrides(entry.fragments.map((f) => f.text));
this.localStore.set(sessionId, entry);
this.logger.debug(
`Set system override for session ${sessionId} (local, ${entry.fragments.length} fragment(s))`,
);
return;
}
// Load existing fragments
const existing = await this.handle.redis.get(SESSION_SYSTEM_FRAGMENTS_KEY(sessionId));
const fragments: OverrideFragment[] = existing
@@ -50,10 +90,17 @@ export class SystemOverrideService {
}
async get(sessionId: string): Promise<string | null> {
if (!this.handle) {
return this.localStore.get(sessionId)?.condensed ?? null;
}
return this.handle.redis.get(SESSION_SYSTEM_KEY(sessionId));
}
async renew(sessionId: string): Promise<void> {
if (!this.handle) {
// Local tier: no TTL to renew; entry persists until restart
return;
}
const pipeline = this.handle.redis.pipeline();
pipeline.expire(SESSION_SYSTEM_KEY(sessionId), SYSTEM_OVERRIDE_TTL_SECONDS);
pipeline.expire(SESSION_SYSTEM_FRAGMENTS_KEY(sessionId), SYSTEM_OVERRIDE_TTL_SECONDS);
@@ -61,6 +108,11 @@ export class SystemOverrideService {
}
async clear(sessionId: string): Promise<void> {
if (!this.handle) {
this.localStore.delete(sessionId);
this.logger.debug(`Cleared system override for session ${sessionId} (local)`);
return;
}
await this.handle.redis.del(
SESSION_SYSTEM_KEY(sessionId),
SESSION_SYSTEM_FRAGMENTS_KEY(sessionId),

View File

@@ -8,7 +8,9 @@ import {
} from '@nestjs/common';
import { Queue, Worker, type Job, type ConnectionOptions } from 'bullmq';
import type { LogService } from '@mosaicstack/log';
import type { MosaicConfig } from '@mosaicstack/config';
import { LOG_SERVICE } from '../log/log.tokens.js';
import { MOSAIC_CONFIG } from '../config/config.module.js';
import type { JobDto, JobStatus } from './queue-admin.dto.js';
// ---------------------------------------------------------------------------
@@ -108,21 +110,42 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
private readonly connection: ConnectionOptions;
private readonly queues = new Map<string, Queue<MosaicJobData>>();
private readonly workers = new Map<string, Worker<MosaicJobData>>();
/** False on Local tier — BullMQ/Redis operations become no-ops. */
private readonly enabled: boolean;
constructor(
@Optional()
@Inject(LOG_SERVICE)
private readonly logService: LogService | null,
@Optional()
@Inject(MOSAIC_CONFIG)
private readonly mosaicConfig: MosaicConfig | null,
) {
this.connection = getConnection();
this.enabled = this.mosaicConfig?.queue?.type !== 'local';
this.connection = this.enabled
? getConnection()
: ({ host: '127.0.0.1', port: 6380 } as ConnectionOptions);
}
/** Returns true when BullMQ/Redis is active (Standalone and Federated tiers). */
isEnabled(): boolean {
return this.enabled;
}
onModuleInit(): void {
this.logger.log('QueueService initialised (BullMQ)');
if (this.enabled) {
this.logger.log('QueueService initialised (BullMQ)');
} else {
this.logger.log(
'QueueService: BullMQ disabled for local tier — no Redis connections will be opened',
);
}
}
async onModuleDestroy(): Promise<void> {
await this.closeAll();
if (this.enabled) {
await this.closeAll();
}
}
// -------------------------------------------------------------------------
@@ -131,8 +154,10 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
/**
* Get or create a BullMQ Queue for the given queue name.
* Returns null on Local tier where BullMQ is disabled.
*/
getQueue<T extends MosaicJobData = MosaicJobData>(name: string): Queue<T> {
getQueue<T extends MosaicJobData = MosaicJobData>(name: string): Queue<T> | null {
if (!this.enabled) return null;
let queue = this.queues.get(name) as Queue<T> | undefined;
if (!queue) {
queue = new Queue<T>(name, { connection: this.connection });
@@ -144,6 +169,7 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
/**
* Add a BullMQ repeatable job (cron-style).
* Uses `jobId` as a deterministic key so duplicate registrations are idempotent.
* No-op on Local tier.
*/
async addRepeatableJob<T extends MosaicJobData>(
queueName: string,
@@ -151,7 +177,13 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
data: T,
cronExpression: string,
): Promise<void> {
const queue = this.getQueue<T>(queueName);
if (!this.enabled) {
this.logger.debug(
`Skipping repeatable job "${jobName}" on "${queueName}" (local tier — BullMQ disabled)`,
);
return;
}
const queue = this.getQueue<T>(queueName)!;
// eslint-disable-next-line @typescript-eslint/no-explicit-any
await (queue as Queue<any>).add(jobName, data, {
repeat: { pattern: cronExpression },
@@ -165,8 +197,18 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
/**
* Register a Worker for the given queue name with error handling and
* exponential backoff.
* Returns null on Local tier where BullMQ is disabled.
*/
registerWorker<T extends MosaicJobData>(queueName: string, handler: JobHandler<T>): Worker<T> {
registerWorker<T extends MosaicJobData>(
queueName: string,
handler: JobHandler<T>,
): Worker<T> | null {
if (!this.enabled) {
this.logger.debug(
`Skipping worker registration for "${queueName}" (local tier — BullMQ disabled)`,
);
return null;
}
const worker = new Worker<T>(
queueName,
async (job) => {
@@ -223,8 +265,12 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
/**
* Return queue health statistics for all managed queues.
* Returns an empty healthy result on Local tier.
*/
async getHealthStatus(): Promise<QueueHealthStatus> {
if (!this.enabled) {
return { queues: {}, healthy: true };
}
const queues: QueueHealthStatus['queues'] = {};
let healthy = true;
@@ -255,8 +301,10 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
/**
* List jobs across all managed queues, optionally filtered by status.
* BullMQ jobs are fetched by state type from each queue.
* Returns empty array on Local tier.
*/
async listJobs(status?: JobStatus): Promise<JobDto[]> {
if (!this.enabled) return [];
const jobs: JobDto[] = [];
const states: JobStatus[] = status
? [status]
@@ -283,8 +331,10 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
* Retry a specific failed job by its BullMQ job ID (format: "queueName:id").
* The caller passes "<queueName>__<jobId>" as the composite ID because BullMQ
* job IDs are not globally unique — they are scoped to their queue.
* Returns an error on Local tier.
*/
async retryJob(compositeId: string): Promise<{ ok: boolean; message: string }> {
if (!this.enabled) return { ok: false, message: 'BullMQ is disabled on local tier.' };
const sep = compositeId.lastIndexOf('__');
if (sep === -1) {
return { ok: false, message: 'Invalid job id format. Expected "<queue>__<jobId>".' };
@@ -316,6 +366,7 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
* Pause a queue by name.
*/
async pauseQueue(name: string): Promise<{ ok: boolean; message: string }> {
if (!this.enabled) return { ok: false, message: 'BullMQ is disabled on local tier.' };
const queue = this.queues.get(name);
if (!queue) return { ok: false, message: `Queue "${name}" not found.` };
await queue.pause();
@@ -327,6 +378,7 @@ export class QueueService implements OnModuleInit, OnModuleDestroy {
* Resume a paused queue by name.
*/
async resumeQueue(name: string): Promise<{ ok: boolean; message: string }> {
if (!this.enabled) return { ok: false, message: 'BullMQ is disabled on local tier.' };
const queue = this.queues.get(name);
if (!queue) return { ok: false, message: `Queue "${name}" not found.` };
await queue.resume();

View File

@@ -91,22 +91,22 @@ Goal: Two federated gateways exchange real data over mTLS. Inbound requests pass
>
> **Tracking issue:** #462.
| id | status | description | issue | agent | branch | depends_on | estimate | notes |
| --------- | ----------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----- | ------ | ------------------------------------ | --------------------------------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
| FED-M3-01 | done | `packages/types/src/federation/` — request/response DTOs for `list`, `get`, `capabilities` verbs. Wire-format zod schemas + inferred TS types. Includes `FederationRequest`, `FederationListResponse<T>`, `FederationGetResponse<T>`, `FederationCapabilitiesResponse`, error envelope, `_source` tag. | #462 | sonnet | feat/federation-m3-types | — | 4K | Reusable from gateway server + client + harness. Pure types — no I/O, no NestJS. |
| FED-M3-02 | done | `tools/federation-harness/` scaffold: `docker-compose.two-gateways.yml` (Server A + Server B + step-CA), `seed.ts` (provisions grants, peers, sample tasks/notes/credentials per scope variant), `harness.ts` helper (boots stack, returns typed clients). README documents harness use. | #462 | sonnet | feat/federation-m3-harness | DEPLOY-04 (soft) | 8K | Falls back to local docker-compose if `mos-test-1/-2` not yet redeployed (DEPLOY chain blocked on IMG-FIX). Permanent test infra used by M3+. |
| FED-M3-03 | done | `apps/gateway/src/federation/server/federation-auth.guard.ts` (NestJS guard). Validates inbound client cert from Fastify TLS context, extracts `grantId` + `subjectUserId` from custom OIDs, loads grant from DB, asserts `status='active'`, attaches `FederationContext` to request. | #462 | sonnet | feat/federation-m3-auth-guard | M3-01 | 8K | Reuses OID parsing logic mirrored from `ca.service.ts` post-issuance verification. 401 on malformed/missing OIDs; 403 on revoked/expired/missing grant. |
| FED-M3-04 | in-progress | `apps/gateway/src/federation/server/scope.service.ts`. Pipeline: (1) resource allowlist + excluded check, (2) native RBAC eval as `subjectUserId`, (3) scope filter intersection (`include_teams`, `include_personal`), (4) `max_rows_per_query` cap. Pure service — DB calls injected. | #462 | sonnet | feat/federation-m3-scope-service | M3-01 | 10K | Hardest correctness target in M3. Reuses `parseFederationScope` (M2-03). Returns either `{ allowed: true, filter }` or structured deny reason for audit. |
| FED-M3-05 | in-progress | `apps/gateway/src/federation/server/verbs/list.controller.ts`. Wires AuthGuard → ScopeService → tasks/notes/memory query layer; applies row cap; tags rows with `_source`. Resource selector via path param. | #462 | sonnet | feat/federation-m3-verb-list | M3-03, M3-04 | 6K | Routes: `POST /api/federation/v1/list/:resource`. No body persistence. Audit write deferred to M4. |
| FED-M3-06 | not-started | `apps/gateway/src/federation/server/verbs/get.controller.ts`. Single-resource fetch by id; same pipeline as list. 404 on not-found, 403 on RBAC/scope deny — both audited the same way. | #462 | sonnet | feat/federation-m3-verb-get | M3-03, M3-04 | 6K | `POST /api/federation/v1/get/:resource/:id`. Mirrors list controller patterns. |
| FED-M3-07 | done | `apps/gateway/src/federation/server/verbs/capabilities.controller.ts`. Read-only enumeration: returns `{ resources, excluded_resources, max_rows_per_query, supported_verbs }` derived from grant scope. Always allowed for an active grant — no RBAC eval. | #462 | sonnet | feat/federation-m3-verb-capabilities | M3-03 | 4K | `GET /api/federation/v1/capabilities`. Smallest verb; useful sanity check that mTLS + auth guard work end-to-end. |
| FED-M3-08 | done | `apps/gateway/src/federation/client/federation-client.service.ts`. Outbound mTLS dialer: picks `(certPem, sealed clientKey)` from `federation_peers`, unwraps key, builds undici Agent with mTLS, calls peer verb, parses typed response, wraps non-2xx into `FederationClientError`. | #462 | sonnet | feat/federation-m3-client | M3-01 | 8K | Independent of server stream — can land in parallel with M3-03/04. Cert/key cached per-peer; flushed by future M5/M6 logic. |
| FED-M3-09 | done | `apps/gateway/src/federation/client/query-source.service.ts`. Accepts `source: "local" \| "federated:<host>" \| "all"` from gateway query layer; for `"all"` fans out to local + each peer in parallel; merges results; tags every row with `_source`. | #462 | sonnet | feat/federation-m3-query-source | M3-08 | 8K | Per-peer failure surfaces as `_partial: true` in response, not hard failure (sets up M5 offline UX). M5 adds caching + circuit breaker on top. |
| FED-M3-10 | not-started | Integration tests for MILESTONES.md M3 acceptance #6 (malformed OIDs → 401; valid cert + revoked grant → 403) and #7 (`max_rows_per_query` cap). Real PG, mocked TLS context (Fastify req shim). | #462 | sonnet | feat/federation-m3-integration | M3-05, M3-06 | 8K | Vitest profile gated by `FEDERATED_INTEGRATION=1`. Single-gateway suite; no harness required. |
| FED-M3-11 | not-started | E2E tests for MILESTONES.md M3 acceptance #1, #2, #3, #4, #5, #8, #9, #10 (8 cases). Uses harness from M3-02; two real gateways, real Step-CA, real mTLS. Each test asserts both happy-path response and audit/no-persist invariants. | #462 | sonnet | feat/federation-m3-e2e | M3-02, M3-04, M3-05, M3-06, M3-09 | 12K | Largest single task. Each acceptance gets its own `it(...)` for clear failure attribution. |
| FED-M3-12 | not-started | Independent security review (sonnet, not author of M3-03/04/05/06/07/08/09): focus on cert-SAN spoofing, OID extraction edge cases, scope-bypass via filter manipulation, RBAC-bypass via subjectUser swap, response leakage when scope deny. | #462 | sonnet | feat/federation-m3-security-review | M3-11 | 10K | Two review rounds budgeted. PRD requires explicit test for every 401/403 path — review verifies coverage. |
| FED-M3-13 | not-started | Docs update: `docs/federation/SETUP.md` mTLS handshake section, new `docs/federation/HARNESS.md` for federation-harness usage, OID reference table in SETUP.md, scope enforcement pipeline diagram. Runbook still M7-deferred. | #462 | haiku | feat/federation-m3-docs | M3-12 | 5K | One ASCII diagram for the auth-guard → scope → RBAC pipeline; helps future reviewers reason about denial paths. |
| FED-M3-14 | not-started | PR aggregate close, CI green, merge to main, close #462. Release tag `fed-v0.3.0-m3`. Update mission manifest M3 row → done; M4 row → in-progress when work begins. | #462 | sonnet | chore/federation-m3-close | M3-13 | 3K | Same close pattern as M1-12 / M2-13. |
| id | status | description | issue | agent | branch | depends_on | estimate | notes |
| --------- | ----------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----- | ------ | ------------------------------------ | ---------------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------- |
| FED-M3-01 | not-started | `packages/types/src/federation/` — request/response DTOs for `list`, `get`, `capabilities` verbs. Wire-format zod schemas + inferred TS types. Includes `FederationRequest`, `FederationListResponse<T>`, `FederationGetResponse<T>`, `FederationCapabilitiesResponse`, error envelope, `_source` tag. | #462 | sonnet | feat/federation-m3-types | — | 4K | Reusable from gateway server + client + harness. Pure types — no I/O, no NestJS. |
| FED-M3-02 | not-started | `tools/federation-harness/` scaffold: `docker-compose.two-gateways.yml` (Server A + Server B + step-CA), `seed.ts` (provisions grants, peers, sample tasks/notes/credentials per scope variant), `harness.ts` helper (boots stack, returns typed clients). README documents harness use. | #462 | sonnet | feat/federation-m3-harness | DEPLOY-04 (soft) | 8K | Falls back to local docker-compose if `mos-test-1/-2` not yet redeployed (DEPLOY chain blocked on IMG-FIX). Permanent test infra used by M3+. |
| FED-M3-03 | not-started | `apps/gateway/src/federation/server/federation-auth.guard.ts` (NestJS guard). Validates inbound client cert from Fastify TLS context, extracts `grantId` + `subjectUserId` from custom OIDs, loads grant from DB, asserts `status='active'`, attaches `FederationContext` to request. | #462 | sonnet | feat/federation-m3-auth-guard | M3-01 | 8K | Reuses OID parsing logic mirrored from `ca.service.ts` post-issuance verification. 401 on malformed/missing OIDs; 403 on revoked/expired/missing grant. |
| FED-M3-04 | not-started | `apps/gateway/src/federation/server/scope.service.ts`. Pipeline: (1) resource allowlist + excluded check, (2) native RBAC eval as `subjectUserId`, (3) scope filter intersection (`include_teams`, `include_personal`), (4) `max_rows_per_query` cap. Pure service — DB calls injected. | #462 | sonnet | feat/federation-m3-scope-service | M3-01 | 10K | Hardest correctness target in M3. Reuses `parseFederationScope` (M2-03). Returns either `{ allowed: true, filter }` or structured deny reason for audit. |
| FED-M3-05 | not-started | `apps/gateway/src/federation/server/verbs/list.controller.ts`. Wires AuthGuard → ScopeService → tasks/notes/memory query layer; applies row cap; tags rows with `_source`. Resource selector via path param. | #462 | sonnet | feat/federation-m3-verb-list | M3-03, M3-04 | 6K | Routes: `POST /api/federation/v1/list/:resource`. No body persistence. Audit write deferred to M4. |
| FED-M3-06 | not-started | `apps/gateway/src/federation/server/verbs/get.controller.ts`. Single-resource fetch by id; same pipeline as list. 404 on not-found, 403 on RBAC/scope deny — both audited the same way. | #462 | sonnet | feat/federation-m3-verb-get | M3-03, M3-04 | 6K | `POST /api/federation/v1/get/:resource/:id`. Mirrors list controller patterns. |
| FED-M3-07 | not-started | `apps/gateway/src/federation/server/verbs/capabilities.controller.ts`. Read-only enumeration: returns `{ resources, excluded_resources, max_rows_per_query, supported_verbs }` derived from grant scope. Always allowed for an active grant — no RBAC eval. | #462 | sonnet | feat/federation-m3-verb-capabilities | M3-03 | 4K | `GET /api/federation/v1/capabilities`. Smallest verb; useful sanity check that mTLS + auth guard work end-to-end. |
| FED-M3-08 | not-started | `apps/gateway/src/federation/client/federation-client.service.ts`. Outbound mTLS dialer: picks `(certPem, sealed clientKey)` from `federation_peers`, unwraps key, builds undici Agent with mTLS, calls peer verb, parses typed response, wraps non-2xx into `FederationClientError`. | #462 | sonnet | feat/federation-m3-client | M3-01 | 8K | Independent of server stream — can land in parallel with M3-03/04. Cert/key cached per-peer; flushed by future M5/M6 logic. |
| FED-M3-09 | not-started | `apps/gateway/src/federation/client/query-source.service.ts`. Accepts `source: "local" \| "federated:<host>" \| "all"` from gateway query layer; for `"all"` fans out to local + each peer in parallel; merges results; tags every row with `_source`. | #462 | sonnet | feat/federation-m3-query-source | M3-08 | 8K | Per-peer failure surfaces as `_partial: true` in response, not hard failure (sets up M5 offline UX). M5 adds caching + circuit breaker on top. |
| FED-M3-10 | not-started | Integration tests for MILESTONES.md M3 acceptance #6 (malformed OIDs → 401; valid cert + revoked grant → 403) and #7 (`max_rows_per_query` cap). Real PG, mocked TLS context (Fastify req shim). | #462 | sonnet | feat/federation-m3-integration | M3-05, M3-06 | 8K | Vitest profile gated by `FEDERATED_INTEGRATION=1`. Single-gateway suite; no harness required. |
| FED-M3-11 | not-started | E2E tests for MILESTONES.md M3 acceptance #1, #2, #3, #4, #5, #8, #9, #10 (8 cases). Uses harness from M3-02; two real gateways, real Step-CA, real mTLS. Each test asserts both happy-path response and audit/no-persist invariants. | #462 | sonnet | feat/federation-m3-e2e | M3-02, M3-09 | 12K | Largest single task. Each acceptance gets its own `it(...)` for clear failure attribution. |
| FED-M3-12 | not-started | Independent security review (sonnet, not author of M3-03/04/05/06/07/08/09): focus on cert-SAN spoofing, OID extraction edge cases, scope-bypass via filter manipulation, RBAC-bypass via subjectUser swap, response leakage when scope deny. | #462 | sonnet | feat/federation-m3-security-review | M3-11 | 10K | Two review rounds budgeted. PRD requires explicit test for every 401/403 path — review verifies coverage. |
| FED-M3-13 | not-started | Docs update: `docs/federation/SETUP.md` mTLS handshake section, new `docs/federation/HARNESS.md` for federation-harness usage, OID reference table in SETUP.md, scope enforcement pipeline diagram. Runbook still M7-deferred. | #462 | haiku | feat/federation-m3-docs | M3-12 | 5K | One ASCII diagram for the auth-guard → scope → RBAC pipeline; helps future reviewers reason about denial paths. |
| FED-M3-14 | not-started | PR aggregate close, CI green, merge to main, close #462. Release tag `fed-v0.3.0-m3`. Update mission manifest M3 row → done; M4 row → in-progress when work begins. | #462 | sonnet | chore/federation-m3-close | M3-13 | 3K | Same close pattern as M1-12 / M2-13. |
**M3 estimate:** ~100K tokens (vs MILESTONES.md 40K — same per-task breakdown pattern as M1/M2: tests, review, and docs split out from implementation cost). Largest milestone in the federation mission.
@@ -118,10 +118,6 @@ Goal: Two federated gateways exchange real data over mTLS. Inbound requests pass
**Test bed fallback:** If `mos-test-1.woltje.com` / `mos-test-2.woltje.com` are still blocked on `FED-M2-DEPLOY-IMG-FIX` when M3-11 is ready to run, the harness's local `docker-compose.two-gateways.yml` is a sufficient stand-in. Production-host validation moves to M7 acceptance suite (PRD AC-12).
**Backlog sync — 2026-06-24 (orchestrator):** Status reconciled against `origin/main` (release 0.0.48). Landed on main: **FED-M3-01** (DTOs, PR #506), **FED-M3-02** (harness scaffold, PR #505), **FED-M3-03** (mTLS auth-guard, PR #509 — CRIT-1/2 + HIGH-1..4 remediated in-PR), **FED-M3-08** (outbound mTLS client, PR #508). With M3-01/03/08 merged, three cards became dependency-clear and were dispatched to the idle coder lane: **FED-M3-04** scope.service → coder0 (`feat/federation-m3-scope-service`); **FED-M3-09** query-source + **FED-M3-07** capabilities verb → coder1 (`feat/federation-m3-query-source` first). Reviewer warmed for the M3 trust-boundary PRs. Remaining blocked-by-DAG: M3-05/06 (await M3-04), M3-10 (await M3-05/06), M3-11 (await M3-09), M3-12→14 (tail). Deploy chain (DEPLOY-IMG-FIX → 03/04) still independent of M3 code — harness local docker-compose fallback covers M3-11.
**Backlog sync #2 — 2026-06-24 (orchestrator):** **FED-M3-09** (query-source) merged via PR #673 and **FED-M3-07** (capabilities) merged via PR #674 — both squash-merged on independent agent review-of-record + green CI (formal Gitea approve unavailable under the shared service account; merge is not gated by the self-approve guard). **FED-M3-05** (list verb) dispatched to coder1 (based on the M3-04 branch, rebase onto main once #672 lands). **FED-M3-04** (scope.service, PR #672) is in review-changes (one include_personal no-leak test outstanding). **DAG fix:** corrected `FED-M3-11` depends_on from `M3-02, M3-09``M3-02, M3-04, M3-05, M3-06, M3-09` — the E2E acceptance cases (#1#5, #8#10) exercise list/get over mTLS, so the server verbs + scope service are hard prerequisites; the original edge set omitted them and caused a premature M3-11 dispatch. Note: M3 read-path invariant for M3-11 is **no-persist + existing enrollment audit only** — read-verb audit-log writes are deferred to M4 (see M3-05/06 notes), so M3-11 must not assert read-audit-log entries.
## Milestone 4 — search + audit + rate limit (FED-M4)
_Deferred. Issue #463._

View File

@@ -1,25 +0,0 @@
# Scratchpad — fleet-personas spec timeout
## Objective
Raise the `@mosaicstack/mosaic` Vitest timeout to 30s at config level so filesystem-backed fleet drift-guard specs (`fleet-personas`, `fleet-profiles`, and siblings) stop false-reding under contended CI.
## Plan
1. Move timeout policy into `packages/mosaic/vitest.config.ts` with `testTimeout: 30_000`.
2. Remove the narrower `fleet-personas.spec.ts` local override so PR #677 fixes the suite class, not one file.
3. Run targeted fleet specs plus typecheck/lint/format gates.
4. Commit, queue guard, push, PR update.
## Evidence
- `pnpm --filter @mosaicstack/mosaic test -- src/commands/fleet-personas.spec.ts` — pass (8 tests; initial narrow fix).
- `pnpm typecheck` — pass (41 tasks; initial narrow fix).
- `pnpm lint` — pass (23 tasks; initial narrow fix).
- `pnpm format:check` — pass after formatting this scratchpad (initial narrow fix).
- Package-wide timeout follow-up:
- `pnpm --filter @mosaicstack/mosaic test -- src/commands/fleet-personas.spec.ts src/commands/fleet-profiles.spec.ts` — pass (24 tests).
- `pnpm --filter @mosaicstack/mosaic test` — pass (44 files / 618 tests).
- `pnpm typecheck` — pass (41 tasks).
- `pnpm lint` — pass (23 tasks).
- `pnpm format:check` — pass.

View File

@@ -1,65 +0,0 @@
# FED-M3-07 — Capabilities Verb Scratchpad
## Objective
Implement `GET /api/federation/v1/capabilities` in `apps/gateway/src/federation/server/verbs/capabilities.controller.ts`.
## Scope
- Add read-only capabilities controller under federation server verbs.
- Use `FederationAuthGuard` only; active grant is sufficient and no native RBAC/scope-service eval runs.
- Response shape: `{ resources, excluded_resources, max_rows_per_query, supported_verbs }` derived from grant scope.
- Register controller in `FederationModule`.
- Unit-test happy path, defaults, no-context guard seam, and invalid scope handling.
## Constraints / assumptions
- Issue: #462.
- Branch: `feat/federation-m3-verb-capabilities` from `origin/main` (`3eeed04e`).
- Depends on M3-03 auth guard; guard attaches `request.federationContext.scope` after active-grant validation.
- ASSUMPTION: `supported_verbs` is the M3 verb set from `@mosaicstack/types` (`list`, `get`, `capabilities`).
- ASSUMPTION: `filters`/`rate_limit` are intentionally omitted for FED-M3-07 because the cards response shape lists only the four required fields.
- Budget: no explicit hard cap from orchestrator; working cap ~4K-8K tokens for card implementation + tests + PR cycle.
## Plan
1. Write controller unit tests first.
2. Implement controller and module registration.
3. Run scoped tests + typecheck/lint/format.
4. Run Codex code/security review and remediate.
5. Commit, queue guard, push, PR via wrapper.
## Progress
- 2026-06-24: Intake complete; fresh worktree created from origin/main.
- 2026-06-24: Added `CapabilitiesController`, registered it in `FederationModule`, and added 5 unit tests.
- 2026-06-24: Code/security reviews passed with no findings.
## Tests run
- `pnpm --filter @mosaicstack/gateway test -- capabilities.controller.spec.ts` — PASS (5 tests).
- `pnpm --filter @mosaicstack/gateway typecheck` — PASS.
- `pnpm --filter @mosaicstack/gateway lint` — PASS.
- `pnpm format:check` — PASS.
- `pnpm typecheck` — PASS (41/41 turbo tasks).
- `pnpm lint` — PASS (23/23 turbo tasks).
- `pnpm test` — FAIL in pre-existing/live-DB integration suite: `apps/gateway/src/__tests__/cross-user-isolation.test.ts` cleanup hit PostgreSQL connection/schema state for the `messages` table. Changed capabilities tests passed; failure is outside FED-M3-07 surface. No `fleet-personas.spec` flake encountered.
## Review evidence
- `~/.config/mosaic/tools/codex/codex-code-review.sh --uncommitted` — PASS/approve, no findings.
- `~/.config/mosaic/tools/codex/codex-security-review.sh --uncommitted` — PASS, risk level none, no findings.
## Risks / blockers
- Full repo `pnpm test` may hit known `fleet-personas.spec` flake per orchestrator; ignore that specific flake if encountered.
- Previous card saw local DB schema issue in `cross-user-isolation.test.ts`; scoped capabilities tests should be authoritative for this surface.
## Acceptance evidence mapping
| Acceptance criterion | Evidence |
| -------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------- |
| GET `/api/federation/v1/capabilities` exists | Route metadata test in `capabilities.controller.spec.ts`; scoped test PASS |
| Uses active-grant auth guard and no RBAC eval | Guard metadata test confirms only `FederationAuthGuard`; controller has no service injections/RBAC calls; scoped test PASS |
| Response enumerates resources/excluded/max rows/supported verbs from scope | Happy-path/default scope tests + response schema parse; scoped test PASS |
| Read-only/no persistence side effects | Controller only parses request `federationContext.scope` and returns a DTO; no DB/service dependency; code review PASS |

View File

@@ -1,67 +0,0 @@
# FED-M3-09 — Query Source Service Scratchpad
## Objective
Implement `apps/gateway/src/federation/client/query-source.service.ts` for `source: "local" | "federated:<host>" | "all"` routing.
## Scope
- Add QuerySourceService in gateway federation client layer.
- Unit-test local-only, single federated peer, all-source fan-out/merge, and per-peer partial failures.
- Keep `docs/federation/TASKS.md` read-only per project agent guidance.
## Constraints / assumptions
- Issue: #462.
- Branch: `feat/federation-m3-query-source` from `origin/main` (`e0e7be70`).
- ASSUMPTION: `federated:<host>` should match active outbound peers by `commonName` first and by `endpointUrl` host/hostname as compatibility fallback; source tags use `peer.commonName` per `@mosaicstack/types` source-tag docs.
- ASSUMPTION: QuerySourceService provides list/fan-out behavior; get/source routing can be layered later because card acceptance says merge rows.
- ASSUMPTION: `source: "all"` cannot safely return a single continuation cursor for multiple sub-sources; any subquery cursor marks the merged response `_partial: true` + `_truncated: true` while omitting `nextCursor`.
- Budget: no explicit hard cap from orchestrator; working cap ~8K-12K tokens for card 1 implementation + tests + PR cycle.
- OpenBrain unavailable: credential loader failed with missing `/home/jarvis/.config/mosaic/credentials.json`; not blocking code delivery.
## Plan
1. Review federation client/types/db patterns.
2. Write unit tests for source behavior.
3. Implement QuerySourceService and export/register it in FederationModule.
4. Run scoped tests, typecheck, lint, format.
5. Run codex uncommitted review and remediate.
6. Commit, queue guard, push, PR via wrapper.
## Progress
- 2026-06-24: Intake complete; using isolated worktree to avoid dirty orchestrator files in original checkout.
- 2026-06-24: Added QuerySourceService, module export, barrel export, and 7 unit tests.
- 2026-06-24: First Codex review found pagination and port-host matching issues; both remediated with tests.
## Tests run
- `pnpm --filter @mosaicstack/gateway test -- query-source.service.spec.ts` — PASS (7 tests).
- `pnpm --filter @mosaicstack/gateway typecheck` — PASS.
- `pnpm --filter @mosaicstack/gateway lint` — PASS.
- `pnpm format:check` — PASS.
- `pnpm typecheck` — PASS (41/41 turbo tasks).
- `pnpm lint` — PASS (23/23 turbo tasks).
- `pnpm test` — FAIL in pre-existing/live-DB integration suite: `apps/gateway/src/__tests__/cross-user-isolation.test.ts` cleanup hit `relation "messages" does not exist` against local PostgreSQL. Changed QuerySource unit tests passed; failure is outside FED-M3-09 surface and appears tied to local DB schema state.
## Review evidence
- `~/.config/mosaic/tools/codex/codex-code-review.sh --uncommitted` — first pass request-changes, 2 should-fix findings (all-source cursor handling; endpoint port host matching).
- Remediation: `_partial` + `_truncated` when any all-source subquery has `nextCursor`; endpoint match accepts URL `host` and `hostname`; added tests for both.
- `~/.config/mosaic/tools/codex/codex-code-review.sh --uncommitted` — PASS/approve, no findings.
- `~/.config/mosaic/tools/codex/codex-security-review.sh --uncommitted` — PASS, risk level none, no findings.
## Risks / blockers
- Federation query layer is not yet wired; service API needs to be stable and easy to compose.
- Must avoid hard-failing `source: all` on remote peer failures.
## Acceptance evidence mapping
| Acceptance criterion | Evidence |
| ------------------------------------------------------------------------------- | --------------------------------------------------------------------------------- |
| local source returns local rows tagged `_source: local` | `query-source.service.spec.ts` local test; scoped test PASS |
| `federated:<host>` queries selected peer and tags rows with peer source | `query-source.service.spec.ts` commonName/endpoint-host tests; scoped test PASS |
| `all` fans out local + active outbound peers in parallel and merges tagged rows | `query-source.service.spec.ts` all-source call-order/merge test; scoped test PASS |
| per-peer failure on `all` returns `_partial: true`, not throw | `query-source.service.spec.ts` peer failure test; scoped test PASS |

View File

@@ -30,7 +30,6 @@ export default tseslint.config(
'apps/gateway/vitest.config.ts',
'packages/db/vitest.config.ts',
'packages/storage/vitest.config.ts',
'packages/mosaic/vitest.config.ts',
'packages/mosaic/__tests__/*.ts',
'tools/federation-harness/*.ts',
],

View File

@@ -69,6 +69,8 @@ describe('Unified wizard (runWizard with default skipGateway)', () => {
const prompter = new HeadlessPrompter({
'Installation mode': 'quick',
'Select your LLM provider': 'anthropic',
'Anthropic API key': 'sk-ant-api03-test',
'What name should agents use?': 'TestBot',
'Communication style': 'direct',
'Your name': 'Tester',
@@ -103,6 +105,8 @@ describe('Unified wizard (runWizard with default skipGateway)', () => {
const prompter = new HeadlessPrompter({
'Installation mode': 'quick',
'Select your LLM provider': 'anthropic',
'Anthropic API key': 'sk-ant-api03-test',
'What name should agents use?': 'TestBot',
'Communication style': 'direct',
'Your name': 'Tester',
@@ -125,6 +129,8 @@ describe('Unified wizard (runWizard with default skipGateway)', () => {
it('respects skipGateway: true', async () => {
const prompter = new HeadlessPrompter({
'Installation mode': 'quick',
'Select your LLM provider': 'anthropic',
'Anthropic API key': 'sk-ant-api03-test',
'What name should agents use?': 'TestBot',
'Communication style': 'direct',
'Your name': 'Tester',

View File

@@ -1,7 +1,7 @@
export class WizardCancelledError extends Error {
override name = 'WizardCancelledError';
constructor() {
super('Wizard cancelled by user');
constructor(message = 'Wizard cancelled by user') {
super(message);
}
}

View File

@@ -207,9 +207,16 @@ export async function finalizeStage(
: 'none selected'
: `install failed — ${skillsResult.failureReason ?? 'unknown error'}`;
const providerConfigured =
state.providerType && state.providerType !== 'none' && state.providerKey;
const providerSummary = providerConfigured
? `Provider: ${state.providerType} (configured)`
: 'Provider: NONE — agent has no brain';
const summary: string[] = [
`Agent: ${state.soul.agentName ?? 'Assistant'}`,
`Style: ${state.soul.communicationStyle ?? 'direct'}`,
providerSummary,
`Runtimes: ${state.runtimes.detected.join(', ') || 'none detected'}`,
`Skills: ${skillsSummary}`,
`Config: ${state.mosaicHome}`,
@@ -239,5 +246,12 @@ export async function finalizeStage(
p.note(nextSteps.map((s, i) => `${(i + 1).toString()}. ${s}`).join('\n'), 'Next Steps');
p.outro('Mosaic is ready.');
if (!providerConfigured) {
p.warn(
'Installation complete, but no LLM provider is configured. ' +
'Run `mosaic wizard` or `mosaic gateway install` to add an API key before using the agent.',
);
} else {
p.outro('Mosaic is ready.');
}
}

View File

@@ -294,7 +294,12 @@ export async function gatewayConfigStage(
}
// Install the gateway npm package on first install or after failure.
if (!opts.skipInstall && !daemonRunning) {
// MOSAIC_GATEWAY_SKIP_NPM_INSTALL=1 forces a skip even without opts.skipInstall:
// used by dev/offline installs where @mosaicstack/gateway is already present
// globally (e.g. a build-from-source `install.sh --dev`) and must not be
// overwritten by the registry @latest build.
const skipNpmInstall = opts.skipInstall || process.env['MOSAIC_GATEWAY_SKIP_NPM_INSTALL'] === '1';
if (!skipNpmInstall && !daemonRunning) {
installGatewayPackage();
}

View File

@@ -78,7 +78,7 @@ describe('providerSetupStage', () => {
expect(state.providerType).toBe('none');
});
it('prompts for key in interactive mode', async () => {
it('prompts for provider then key in interactive mode', async () => {
delete process.env['MOSAIC_ASSUME_YES'];
// Simulate a TTY
const origIsTTY = process.stdin.isTTY;
@@ -86,11 +86,13 @@ describe('providerSetupStage', () => {
const state = makeState();
const p = buildPrompter({
select: vi.fn().mockResolvedValue('anthropic'),
text: vi.fn().mockResolvedValue('sk-ant-api03-interactive'),
});
await providerSetupStage(p, state);
expect(p.select).toHaveBeenCalled();
expect(p.text).toHaveBeenCalled();
expect(state.providerKey).toBe('sk-ant-api03-interactive');
expect(state.providerType).toBe('anthropic');
@@ -98,20 +100,57 @@ describe('providerSetupStage', () => {
Object.defineProperty(process.stdin, 'isTTY', { value: origIsTTY, configurable: true });
});
it('handles empty key in interactive mode', async () => {
it('rejects empty and mismatched keys via the validate callback (Anthropic)', async () => {
delete process.env['MOSAIC_ASSUME_YES'];
const origIsTTY = process.stdin.isTTY;
Object.defineProperty(process.stdin, 'isTTY', { value: true, configurable: true });
let capturedValidate: ((v: string) => string | void) | undefined;
const state = makeState();
const p = buildPrompter({
text: vi.fn().mockResolvedValue(''),
select: vi.fn().mockResolvedValue('anthropic'),
text: vi
.fn()
.mockImplementation(async (opts: { validate?: (v: string) => string | void }) => {
capturedValidate = opts.validate;
return 'sk-ant-api03-ok';
}),
});
await providerSetupStage(p, state);
expect(state.providerType).toBe('none');
expect(state.providerKey).toBeUndefined();
expect(capturedValidate).toBeDefined();
expect(capturedValidate?.('')).toBe('API key is required');
expect(capturedValidate?.(' ')).toBe('API key is required');
expect(capturedValidate?.('not-a-key')).toBe('Anthropic keys start with sk-ant-');
expect(capturedValidate?.('sk-ant-valid')).toBeUndefined();
expect(state.providerType).toBe('anthropic');
Object.defineProperty(process.stdin, 'isTTY', { value: origIsTTY, configurable: true });
});
it('rejects an Anthropic key when OpenAI is selected', async () => {
delete process.env['MOSAIC_ASSUME_YES'];
const origIsTTY = process.stdin.isTTY;
Object.defineProperty(process.stdin, 'isTTY', { value: true, configurable: true });
let capturedValidate: ((v: string) => string | void) | undefined;
const state = makeState();
const p = buildPrompter({
select: vi.fn().mockResolvedValue('openai'),
text: vi
.fn()
.mockImplementation(async (opts: { validate?: (v: string) => string | void }) => {
capturedValidate = opts.validate;
return 'sk-proj-ok';
}),
});
await providerSetupStage(p, state);
expect(capturedValidate?.('sk-ant-api03-xyz')).toBe('OpenAI keys start with sk- (not sk-ant-)');
expect(capturedValidate?.('sk-proj-xyz')).toBeUndefined();
expect(state.providerType).toBe('openai');
Object.defineProperty(process.stdin, 'isTTY', { value: origIsTTY, configurable: true });
});

View File

@@ -1,12 +1,13 @@
import type { WizardPrompter } from '../prompter/interface.js';
import type { WizardState } from '../types.js';
import { detectProviderType } from '../constants.js';
import type { ProviderType } from '../types.js';
/**
* Provider setup stage — collects the user's LLM API key and detects the
* Provider setup stage — collects the user's LLM API key and validates the
* provider type from the key prefix.
*
* In headless mode, reads from `MOSAIC_ANTHROPIC_API_KEY` or `MOSAIC_OPENAI_API_KEY`.
* Interactive mode requires the user to select a provider and enter a valid key.
*/
export async function providerSetupStage(p: WizardPrompter, state: WizardState): Promise<void> {
const isHeadless = process.env['MOSAIC_ASSUME_YES'] === '1' || !process.stdin.isTTY;
@@ -16,39 +17,57 @@ export async function providerSetupStage(p: WizardPrompter, state: WizardState):
const openaiKey = process.env['MOSAIC_OPENAI_API_KEY'] ?? '';
const key = anthropicKey || openaiKey;
state.providerKey = key || undefined;
state.providerType = detectProviderType(key);
if (anthropicKey) {
state.providerType = 'anthropic';
} else if (openaiKey) {
state.providerType = 'openai';
} else {
state.providerType = 'none';
p.warn(
'No API key found (MOSAIC_ANTHROPIC_API_KEY / MOSAIC_OPENAI_API_KEY). ' +
'Run `mosaic gateway install` to configure a key before using the agent.',
);
}
return;
}
p.separator();
p.note(
'Configure your LLM provider so the agent has a brain.\n' +
'Anthropic (Claude) and OpenAI are supported.\n' +
'You can skip this and add a key later via `mosaic configure`.',
'Anthropic (Claude) and OpenAI are supported. You will need an API key to continue.',
'LLM Provider',
);
const key = await p.text({
message: 'API key (paste your Anthropic or OpenAI key, or press Enter to skip)',
defaultValue: '',
placeholder: 'sk-ant-api03-... or sk-...',
const providerType = await p.select<ProviderType>({
message: 'Select your LLM provider',
options: [
{ value: 'anthropic', label: 'Anthropic (Claude)', hint: 'Keys start with sk-ant-' },
{ value: 'openai', label: 'OpenAI', hint: 'Keys start with sk-' },
],
initialValue: 'anthropic',
});
if (key) {
const provider = detectProviderType(key);
state.providerKey = key;
state.providerType = provider;
const key = await p.text({
message: providerType === 'anthropic' ? 'Anthropic API key' : 'OpenAI API key',
placeholder: providerType === 'anthropic' ? 'sk-ant-api03-...' : 'sk-...',
validate: (value: string): string | void => {
if (!value || value.trim().length === 0) {
return 'API key is required';
}
const trimmed = value.trim();
if (providerType === 'anthropic' && !trimmed.startsWith('sk-ant-')) {
return 'Anthropic keys start with sk-ant-';
}
if (
providerType === 'openai' &&
(!trimmed.startsWith('sk-') || trimmed.startsWith('sk-ant-'))
) {
return 'OpenAI keys start with sk- (not sk-ant-)';
}
},
});
if (provider === 'anthropic') {
p.log('Detected provider: Anthropic (Claude)');
} else if (provider === 'openai') {
p.log('Detected provider: OpenAI');
} else {
p.log('Provider auto-detection failed. Key will be stored as ANTHROPIC_API_KEY.');
state.providerType = 'anthropic';
}
} else {
state.providerType = 'none';
p.log('No API key provided. You can add one later with `mosaic configure`.');
}
state.providerKey = key.trim();
state.providerType = providerType;
p.log(`Provider configured: ${providerType === 'anthropic' ? 'Anthropic (Claude)' : 'OpenAI'}`);
}

View File

@@ -2,6 +2,7 @@ import type { WizardPrompter } from '../prompter/interface.js';
import type { ConfigService } from '../config/config-service.js';
import type { WizardState } from '../types.js';
import { DEFAULTS } from '../constants.js';
import { WizardCancelledError } from '../errors.js';
import { providerSetupStage } from './provider-setup.js';
import { runtimeSetupStage } from './runtime-setup.js';
import { hooksPreviewStage } from './hooks-preview.js';
@@ -38,6 +39,25 @@ export async function quickStartPath(
// 1. Provider setup (first question)
await providerSetupStage(prompter, state);
// Belt-and-suspenders guard: ensure a provider key was set before proceeding.
// The interactive path in providerSetupStage always requires a key, so this
// guard is effectively unreachable interactively. The headless path may
// produce providerType='none' when no env var is present: there we warn (the
// operator can configure a key later via `mosaic gateway install`) and let
// the scripted install continue — finalize.ts will NOT print "Mosaic is
// ready" without a configured provider, so no false-green is possible.
if (state.providerType === 'none' || !state.providerKey) {
const headlessRun = process.env['MOSAIC_ASSUME_YES'] === '1' || !process.stdin.isTTY;
if (!headlessRun) {
prompter.warn(
'A provider API key is required to continue. ' +
'Set MOSAIC_ANTHROPIC_API_KEY or MOSAIC_OPENAI_API_KEY and run the wizard again, ' +
'or run `mosaic gateway install` to configure one after installation.',
);
throw new WizardCancelledError('No LLM provider configured');
}
}
// Apply sensible defaults for everything else
state.soul.agentName ??= 'Mosaic';
state.soul.roleDescription ??= DEFAULTS.roleDescription;

View File

@@ -4,6 +4,5 @@ export default defineConfig({
test: {
globals: true,
environment: 'node',
testTimeout: 30_000,
},
});