feat: add LlamaCppClient with basic chat support

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

src/models/local/llamacpp.test.ts

@@ -0,0 +1,36 @@
+import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
+import { LlamaCppClient } from './llamacpp.js';
+
+describe('LlamaCppClient', () => {
+  const mockFetch = vi.fn();
+
+  beforeEach(() => {
+    vi.stubGlobal('fetch', mockFetch);
+  });
+
+  afterEach(() => {
+    vi.unstubAllGlobals();
+  });
+
+  it('sends messages and returns response', async () => {
+    mockFetch.mockResolvedValue({
+      ok: true,
+      json: () => Promise.resolve({
+        choices: [{ message: { content: 'Hello from llama.cpp!' } }],
+        usage: { prompt_tokens: 10, completion_tokens: 5 },
+      }),
+    });
+
+    const client = new LlamaCppClient({
+      endpoint: 'http://localhost:8080',
+    });
+
+    const response = await client.chat({
+      messages: [{ role: 'user', content: 'Hello' }],
+    });
+
+    expect(response.content).toBe('Hello from llama.cpp!');
+    expect(response.usage.inputTokens).toBe(10);
+    expect(response.usage.outputTokens).toBe(5);
+  });
+});