feat: add heartbeat monitor and vector memory search (Tier 2)

Heartbeat:
- HeartbeatMonitor with 5 checks: gateway, model, channels, memory, disk
- Configurable interval, failure threshold, notification channel
- Recovery notifications when health restores
- 25 new tests

Vector Memory Search:
- EmbeddingProvider interface with OpenAI, Gemini, Ollama, LlamaCpp backends
- SQLite-backed VectorStore with cosine similarity search
- Text chunker with paragraph-aware splitting and overlap
- HybridSearch merging keyword + vector results with configurable weight
- Background indexer with dirty-namespace tracking
- Graceful fallback to keyword search when embeddings unavailable
- 51 new tests

Config: automation.heartbeat + memory.embedding schema sections
Total: 950 tests passing, all types clean

src/memory/embeddings.test.ts

@@ -0,0 +1,159 @@
+import { describe, it, expect } from 'vitest';
+import {
+  createEmbeddingProvider,
+  OpenAIEmbeddingProvider,
+  GeminiEmbeddingProvider,
+  OllamaEmbeddingProvider,
+  LlamaCppEmbeddingProvider,
+} from './embeddings.js';
+import type { EmbeddingConfig } from '../config/schema.js';
+
+describe('createEmbeddingProvider', () => {
+  const baseConfig: EmbeddingConfig = {
+    enabled: true,
+    provider: 'openai',
+    model: 'text-embedding-3-small',
+    chunk_size: 512,
+    chunk_overlap: 50,
+    top_k: 5,
+    hybrid_weight: 0.7,
+  };
+
+  it('creates OpenAI provider', () => {
+    const provider = createEmbeddingProvider({ ...baseConfig, provider: 'openai' });
+    expect(provider).toBeInstanceOf(OpenAIEmbeddingProvider);
+  });
+
+  it('creates Gemini provider', () => {
+    const provider = createEmbeddingProvider({ ...baseConfig, provider: 'gemini' });
+    expect(provider).toBeInstanceOf(GeminiEmbeddingProvider);
+  });
+
+  it('creates Ollama provider', () => {
+    const provider = createEmbeddingProvider({ ...baseConfig, provider: 'ollama' });
+    expect(provider).toBeInstanceOf(OllamaEmbeddingProvider);
+  });
+
+  it('creates LlamaCpp provider', () => {
+    const provider = createEmbeddingProvider({ ...baseConfig, provider: 'llamacpp' });
+    expect(provider).toBeInstanceOf(LlamaCppEmbeddingProvider);
+  });
+
+  it('throws on unknown provider', () => {
+    expect(() => createEmbeddingProvider({ ...baseConfig, provider: 'unknown' as never })).toThrow('Unknown embedding provider');
+  });
+});
+
+describe('OpenAIEmbeddingProvider', () => {
+  it('reports configured dimensions', () => {
+    const config: EmbeddingConfig = {
+      enabled: true,
+      provider: 'openai',
+      model: 'text-embedding-3-small',
+      dimensions: 512,
+      chunk_size: 512,
+      chunk_overlap: 50,
+      top_k: 5,
+      hybrid_weight: 0.7,
+    };
+    const provider = new OpenAIEmbeddingProvider(config);
+    expect(provider.dimensions).toBe(512);
+  });
+
+  it('defaults to 1536 dimensions', () => {
+    const config: EmbeddingConfig = {
+      enabled: true,
+      provider: 'openai',
+      model: 'text-embedding-3-small',
+      chunk_size: 512,
+      chunk_overlap: 50,
+      top_k: 5,
+      hybrid_weight: 0.7,
+    };
+    const provider = new OpenAIEmbeddingProvider(config);
+    expect(provider.dimensions).toBe(1536);
+  });
+});
+
+describe('GeminiEmbeddingProvider', () => {
+  it('reports configured dimensions', () => {
+    const config: EmbeddingConfig = {
+      enabled: true,
+      provider: 'gemini',
+      model: 'text-embedding-004',
+      dimensions: 256,
+      chunk_size: 512,
+      chunk_overlap: 50,
+      top_k: 5,
+      hybrid_weight: 0.7,
+    };
+    const provider = new GeminiEmbeddingProvider(config);
+    expect(provider.dimensions).toBe(256);
+  });
+
+  it('defaults to 768 dimensions', () => {
+    const config: EmbeddingConfig = {
+      enabled: true,
+      provider: 'gemini',
+      model: 'text-embedding-004',
+      chunk_size: 512,
+      chunk_overlap: 50,
+      top_k: 5,
+      hybrid_weight: 0.7,
+    };
+    const provider = new GeminiEmbeddingProvider(config);
+    expect(provider.dimensions).toBe(768);
+  });
+});
+
+describe('OllamaEmbeddingProvider', () => {
+  it('reports configured dimensions', () => {
+    const config: EmbeddingConfig = {
+      enabled: true,
+      provider: 'ollama',
+      model: 'nomic-embed-text',
+      dimensions: 384,
+      endpoint: 'http://localhost:11434',
+      chunk_size: 512,
+      chunk_overlap: 50,
+      top_k: 5,
+      hybrid_weight: 0.7,
+    };
+    const provider = new OllamaEmbeddingProvider(config);
+    expect(provider.dimensions).toBe(384);
+  });
+});
+
+describe('LlamaCppEmbeddingProvider', () => {
+  it('reports configured dimensions', () => {
+    const config: EmbeddingConfig = {
+      enabled: true,
+      provider: 'llamacpp',
+      model: 'unused',
+      dimensions: 768,
+      endpoint: 'http://localhost:8080',
+      chunk_size: 512,
+      chunk_overlap: 50,
+      top_k: 5,
+      hybrid_weight: 0.7,
+    };
+    const provider = new LlamaCppEmbeddingProvider(config);
+    expect(provider.dimensions).toBe(768);
+  });
+
+  it('defaults endpoint to localhost:8080', () => {
+    const config: EmbeddingConfig = {
+      enabled: true,
+      provider: 'llamacpp',
+      model: 'unused',
+      dimensions: 768,
+      chunk_size: 512,
+      chunk_overlap: 50,
+      top_k: 5,
+      hybrid_weight: 0.7,
+    };
+    // Provider should be constructable without endpoint
+    const provider = new LlamaCppEmbeddingProvider(config);
+    expect(provider.dimensions).toBe(768);
+  });
+});