feat(llm-gateway): implement LLM Gateway MCP Server (#56)

Implements complete LLM Gateway MCP Server with:
- FastMCP server with 4 tools: chat_completion, list_models, get_usage, count_tokens
- LiteLLM Router with multi-provider failover chains
- Circuit breaker pattern for fault tolerance
- Redis-based cost tracking per project/agent
- Comprehensive test suite (209 tests, 92% coverage)

Model groups defined per ADR-004:
- reasoning: claude-opus-4 → gpt-4.1 → gemini-2.5-pro
- code: claude-sonnet-4 → gpt-4.1 → deepseek-coder
- fast: claude-haiku → gpt-4.1-mini → gemini-2.0-flash

🤖 Generated with [Claude Code](https://claude.com/claude-code)

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

mcp-servers/llm-gateway/pyproject.toml

@@ -4,20 +4,96 @@ version = "0.1.0"
 description = "Syndarix LLM Gateway MCP Server - Unified LLM access with failover and cost tracking"
 requires-python = ">=3.12"
 dependencies = [
-    "fastmcp>=0.1.0",
+    "fastmcp>=2.0.0",
     "litellm>=1.50.0",
     "redis>=5.0.0",
     "pydantic>=2.0.0",
     "pydantic-settings>=2.0.0",
+    "tiktoken>=0.7.0",
+    "httpx>=0.27.0",
+    "uvicorn>=0.30.0",
+    "fastapi>=0.115.0",
 ]
 
 [project.optional-dependencies]
 dev = [
     "pytest>=8.0.0",
-    "pytest-asyncio>=0.23.0",
+    "pytest-asyncio>=0.24.0",
+    "pytest-cov>=5.0.0",
+    "respx>=0.21.0",
+    "fakeredis>=2.25.0",
     "ruff>=0.8.0",
+    "mypy>=1.11.0",
 ]
 
+[project.scripts]
+llm-gateway = "server:main"
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["."]
+exclude = ["tests/", "*.md", "Dockerfile"]
+
+[tool.hatch.build.targets.sdist]
+include = ["*.py", "pyproject.toml"]
+
 [tool.ruff]
 target-version = "py312"
 line-length = 88
+
+[tool.ruff.lint]
+select = [
+    "E",      # pycodestyle errors
+    "W",      # pycodestyle warnings
+    "F",      # pyflakes
+    "I",      # isort
+    "B",      # flake8-bugbear
+    "C4",     # flake8-comprehensions
+    "UP",     # pyupgrade
+    "ARG",    # flake8-unused-arguments
+    "SIM",    # flake8-simplify
+]
+ignore = [
+    "E501",   # line too long (handled by formatter)
+    "B008",   # do not perform function calls in argument defaults
+    "B904",   # raise from in except (too noisy)
+]
+
+[tool.ruff.lint.isort]
+known-first-party = ["config", "models", "exceptions", "providers", "failover", "routing", "cost_tracking", "streaming"]
+
+[tool.pytest.ini_options]
+asyncio_mode = "auto"
+asyncio_default_fixture_loop_scope = "function"
+testpaths = ["tests"]
+addopts = "-v --tb=short"
+filterwarnings = [
+    "ignore::DeprecationWarning",
+]
+
+[tool.coverage.run]
+source = ["."]
+omit = ["tests/*", "conftest.py"]
+branch = true
+
+[tool.coverage.report]
+exclude_lines = [
+    "pragma: no cover",
+    "def __repr__",
+    "raise NotImplementedError",
+    "if TYPE_CHECKING:",
+    "if __name__ == .__main__.:",
+]
+fail_under = 90
+show_missing = true
+
+[tool.mypy]
+python_version = "3.12"
+strict = true
+warn_return_any = true
+warn_unused_ignores = true
+disallow_untyped_defs = true
+plugins = ["pydantic.mypy"]