feat(opencode): remplace le provider Ollama par llama.cpp

Le tool-calling local ne fonctionnait jamais via Ollama. Refonte du
support local d'OpenCode autour de llama.cpp: profil, catalogue,
matérialisation de la config OpenCode et surface first-run alignés sur
llama-server (backend + frontend).

QA vert (commandes réelles): domain 244, application 81+64, infra 263
(10 échecs = bind-port sandbox identiques sur develop, non-régression),
frontend 574, tsc propre. Réserve E2E live non bloquante faute de
llama-server joignable.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

frontend/src/adapters/mock/index.ts

@@ -1160,20 +1160,18 @@ export const MOCK_REFERENCE_PROFILES: AgentProfile[] = [
     cwdTemplate: "{projectRoot}",
   },
   {
-    id: "mock-ollama",
-    name: "Ollama / OpenAI-compatible local model",
-    command: "openai-compatible",
+    id: "mock-opencode",
+    name: "OpenCode + llama.cpp",
+    command: "opencode",
     args: [],
     contextInjection: { strategy: "conventionFile", target: "AGENTS.md" },
-    detect: null,
+    detect: "opencode --version",
     cwdTemplate: "{projectRoot}",
-    structuredAdapter: "openAiCompatible",
-    chatHttp: {
-      endpoint: "http://localhost:11434/v1",
-      model: "qwen2.5-coder",
-      requestTimeoutMs: 120_000,
-      connectTimeoutMs: 5_000,
-      maxToolIterations: 16,
+    structuredAdapter: "openCode",
+    opencode: {
+      baseURL: "http://localhost:8080/v1",
+      apiKey: "sk-no-key",
+      model: "qwen3-coder-30b",
     },
   },
 ];