feat(llm): add Gemini CLI provider as primary; set qwen3.5:9b as default Ollama model

- Add GeminiCliProvider: shells out to `gemini -p` with --output-format json,
  injection-safe prompt passing, MCP server suppression via temp workdir,
  6-slot concurrency semaphore, 60s subprocess deadline
- Add --llm-provider, --llm-model, --llm-base-url CLI flags for per-call overrides
- Provider chain: Gemini CLI → OpenAI → Ollama → Anthropic
- Move LLM timing to dispatch layer (LLM: Xs on stderr)
- Default Ollama model: qwen3:8b → qwen3.5:9b (benchmark shows better schema extraction)
- Add noxa mcp subcommand
- Add docs/reports/llm-benchmark-2026-04-11.md (Gemini vs qwen3.5:4b vs qwen3.5:9b)
- Bump version 0.3.11 → 0.4.0

Co-authored-by: Claude <claude@anthropic.com>

setup.sh

@@ -3,6 +3,8 @@
 #
 # Checks prerequisites, builds binaries, configures .env,
 # optionally installs Ollama, and wires up the MCP server.
+# The generated .env is broader than env.example: it includes local
+# deployment and Ollama settings used by the setup script and compose stack.
 #
 # Usage:
 #   ./setup.sh              # Interactive full setup
@@ -214,6 +216,8 @@ configure_env() {
     fi
 
     # Write .env
+    # env.example covers the runtime noxa variables; this file adds local
+    # deployment and Ollama settings used by setup.sh and docker-compose.yml.
     cat > "$SCRIPT_DIR/.env" <<EOF
 # noxa configuration — generated by setup.sh