Merge pull request #834 from coleam00/security/remove-docker-socket-risk

Security: Remove Docker socket mounting to eliminate CVE-2025-9074 risk

.env.example

@@ -80,6 +80,29 @@ SERVICE_DISCOVERY_MODE=local
 STATE_STORAGE_TYPE=file
 FILE_STATE_DIRECTORY=agent-work-orders-state
 
+# MCP Server Monitoring (Security Configuration)
+# Controls how archon-server monitors MCP server status
+#
+# HTTP Mode (Recommended - Default):
+# - Secure: No Docker socket access required
+# - Portable: Works in Docker, Kubernetes, bare metal
+# - Set: ENABLE_DOCKER_SOCKET_MONITORING=false (or leave unset)
+#
+# Docker Socket Mode (Legacy - Security Risk):
+# - Requires: Docker socket mounted (root-equivalent host access)
+# - Security Risk: CVE-2025-9074 demonstrates container escape vulnerabilities
+# - Only use if: You specifically need Docker container uptime details
+# - Set: ENABLE_DOCKER_SOCKET_MONITORING=true
+# - Also requires: Uncommenting Docker socket volume in docker-compose.yml (line 36)
+#
+# Default: false (HTTP mode, secure)
+ENABLE_DOCKER_SOCKET_MONITORING=false
+
+# MCP Health Check Timeout (seconds)
+# Timeout for HTTP health check requests to MCP server
+# Default: 5
+MCP_HEALTH_CHECK_TIMEOUT=5
+
 # Frontend Configuration
 # VITE_ALLOWED_HOSTS: Comma-separated list of additional hosts allowed for Vite dev server
 # Example: VITE_ALLOWED_HOSTS=192.168.1.100,myhost.local,example.com

docker-compose.yml

@@ -33,7 +33,12 @@ services:
     networks:
       - app-network
     volumes:
-      - /var/run/docker.sock:/var/run/docker.sock # Docker socket for MCP container control
+      # SECURITY: Docker socket mounting removed (CVE-2025-9074 - CVSS 9.3)
+      # MCP status now monitored via HTTP health checks (secure, portable)
+      # To re-enable Docker socket mode (not recommended):
+      #   1. Set ENABLE_DOCKER_SOCKET_MONITORING=true in .env
+      #   2. Uncomment the line below
+      # - /var/run/docker.sock:/var/run/docker.sock # SECURITY RISK: root-equivalent host access
       - ./python/src:/app/src # Mount source code for hot reload
       - ./python/tests:/app/tests # Mount tests for UI test execution
       - ./migration:/app/migration # Mount migration files for version tracking

python/pyproject.toml

@@ -58,7 +58,9 @@ server = [
     "httpx>=0.24.0",
     "pydantic>=2.0.0",
     "python-dotenv>=1.0.0",
-    "docker>=6.1.0",
+    # OPTIONAL: Docker SDK only needed for legacy Docker socket monitoring mode
+    # Uncomment if ENABLE_DOCKER_SOCKET_MONITORING=true (not recommended - security risk)
+    # "docker>=6.1.0",
     "tldextract>=5.0.0",
     # Logging
     "logfire>=0.30.0",
@@ -128,7 +130,9 @@ all = [
     "python-jose[cryptography]>=3.3.0",
     "cryptography>=41.0.0",
     "slowapi>=0.1.9",
-    "docker>=6.1.0",
+    # OPTIONAL: Docker SDK only needed for legacy Docker socket monitoring mode
+    # Uncomment if ENABLE_DOCKER_SOCKET_MONITORING=true (not recommended - security risk)
+    # "docker>=6.1.0",
     "tldextract>=5.0.0",
     "logfire>=0.30.0",
     # MCP specific (mcp version)

python/src/server/api_routes/mcp_api.py

@@ -3,23 +3,93 @@ MCP API endpoints for Archon
 
 Provides status and configuration endpoints for the MCP service.
 The MCP container is managed by docker-compose, not by this API.
+
+Status monitoring uses HTTP health checks by default (secure, portable).
+Docker socket mode available via ENABLE_DOCKER_SOCKET_MONITORING (legacy, security risk).
 """
 
 import os
 from typing import Any
 
-import docker
-from docker.errors import NotFound
+import httpx
 from fastapi import APIRouter, HTTPException
 
 # Import unified logging
+from ..config.config import get_mcp_monitoring_config
 from ..config.logfire_config import api_logger, safe_set_attribute, safe_span
+from ..config.service_discovery import get_mcp_url
 
 router = APIRouter(prefix="/api/mcp", tags=["mcp"])
 
 
-def get_container_status() -> dict[str, Any]:
-    """Get simple MCP container status without Docker management."""
+async def get_container_status_http() -> dict[str, Any]:
+    """Get MCP server status via HTTP health check endpoint.
+
+    This is the secure, recommended approach that doesn't require Docker socket.
+    Works across all deployment environments (Docker, Kubernetes, bare metal).
+
+    Returns:
+        Status dict: {"status": str, "uptime": int|None, "logs": []}
+    """
+    config = get_mcp_monitoring_config()
+    mcp_url = get_mcp_url()
+
+    try:
+        # Use async context manager for proper connection cleanup
+        async with httpx.AsyncClient(timeout=config.health_check_timeout) as client:
+            response = await client.get(f"{mcp_url}/health")
+            response.raise_for_status()
+
+            # MCP health endpoint returns: {"success": bool, "uptime_seconds": int, "health": {...}}
+            data = response.json()
+
+            # Transform to expected API contract
+            uptime_value = data.get("uptime_seconds")
+            return {
+                "status": "running" if data.get("success") else "unhealthy",
+                "uptime": int(uptime_value) if uptime_value is not None else None,
+                "logs": [],  # Historical artifact, kept for API compatibility
+            }
+
+    except httpx.ConnectError:
+        # MCP container not running or unreachable
+        api_logger.warning("MCP server unreachable via HTTP health check")
+        return {
+            "status": "unreachable",
+            "uptime": None,
+            "logs": [],
+        }
+    except httpx.TimeoutException:
+        # MCP responding too slowly
+        api_logger.warning(f"MCP server health check timed out after {config.health_check_timeout}s")
+        return {
+            "status": "unhealthy",
+            "uptime": None,
+            "logs": [],
+        }
+    except Exception:
+        # Unexpected error
+        api_logger.error("Failed to check MCP server health via HTTP", exc_info=True)
+        return {
+            "status": "error",
+            "uptime": None,
+            "logs": [],
+        }
+
+
+def get_container_status_docker() -> dict[str, Any]:
+    """Get MCP container status via Docker socket (legacy mode).
+
+    SECURITY WARNING: Requires Docker socket mounted, granting root-equivalent host access.
+    Only enable this mode if you specifically need Docker container status details.
+    Set ENABLE_DOCKER_SOCKET_MONITORING=true to use this mode.
+
+    Returns:
+        Status dict: {"status": str, "uptime": int|None, "logs": []}
+    """
+    import docker
+    from docker.errors import NotFound
+
     docker_client = None
     try:
         docker_client = docker.from_env()
@@ -34,6 +104,7 @@ def get_container_status() -> dict[str, Any]:
             # Try to get uptime from container info
             try:
                 from datetime import datetime
+
                 started_at = container.attrs["State"]["StartedAt"]
                 started_time = datetime.fromisoformat(started_at.replace("Z", "+00:00"))
                 uptime = int((datetime.now(started_time.tzinfo) - started_time).total_seconds())
@@ -47,27 +118,26 @@ def get_container_status() -> dict[str, Any]:
             "status": status,
             "uptime": uptime,
             "logs": [],  # No log streaming anymore
-            "container_status": container_status
         }
 
     except NotFound:
+        api_logger.warning("MCP container not found via Docker socket")
         return {
             "status": "not_found",
             "uptime": None,
             "logs": [],
-            "container_status": "not_found",
-            "message": "MCP container not found. Run: docker compose up -d archon-mcp"
+            "message": "MCP container not found. Run: docker compose up -d archon-mcp",
         }
     except Exception as e:
-        api_logger.error("Failed to get container status", exc_info=True)
+        api_logger.error("Failed to get MCP container status via Docker", exc_info=True)
         return {
             "status": "error",
             "uptime": None,
             "logs": [],
-            "container_status": "error",
-            "error": str(e)
+            "error": str(e),
         }
     finally:
+        # CRITICAL: Always close Docker client to prevent connection leaks
         if docker_client is not None:
             try:
                 docker_client.close()
@@ -75,15 +145,38 @@ def get_container_status() -> dict[str, Any]:
                 pass
 
 
+async def get_container_status() -> dict[str, Any]:
+    """Get MCP server status using configured monitoring strategy.
+
+    Routes to HTTP health check (secure, default) or Docker socket (legacy).
+
+    Returns:
+        Status dict: {"status": str, "uptime": int|None, "logs": []}
+    """
+    config = get_mcp_monitoring_config()
+
+    if config.enable_docker_socket:
+        api_logger.info("Using Docker socket monitoring (ENABLE_DOCKER_SOCKET_MONITORING=true)")
+        # Docker mode is synchronous
+        return get_container_status_docker()
+    else:
+        # HTTP mode is asynchronous (default)
+        return await get_container_status_http()
+
+
 @router.get("/status")
 async def get_status():
-    """Get MCP server status."""
+    """Get MCP server status.
+
+    Returns container/server status, uptime, and logs (empty).
+    Monitoring strategy controlled by ENABLE_DOCKER_SOCKET_MONITORING env var.
+    """
     with safe_span("api_mcp_status") as span:
         safe_set_attribute(span, "endpoint", "/api/mcp/status")
         safe_set_attribute(span, "method", "GET")
 
         try:
-            status = get_container_status()
+            status = await get_container_status()
             api_logger.debug(f"MCP server status checked - status={status.get('status')}")
             safe_set_attribute(span, "status", status.get("status"))
             safe_set_attribute(span, "uptime", status.get("uptime"))
@@ -91,7 +184,7 @@ async def get_status():
         except Exception as e:
             api_logger.error(f"MCP server status API failed - error={str(e)}")
             safe_set_attribute(span, "error", str(e))
-            raise HTTPException(status_code=500, detail=str(e))
+            raise HTTPException(status_code=500, detail=str(e)) from e
 
 
 @router.get("/config")
@@ -118,9 +211,7 @@ async def get_mcp_config():
             try:
                 from ..services.credential_service import credential_service
 
-                model_choice = await credential_service.get_credential(
-                    "MODEL_CHOICE", "gpt-4o-mini"
-                )
+                model_choice = await credential_service.get_credential("MODEL_CHOICE", "gpt-4o-mini")
                 config["model_choice"] = model_choice
             except Exception:
                 # Fallback to default model
@@ -136,7 +227,7 @@ async def get_mcp_config():
         except Exception as e:
             api_logger.error("Failed to get MCP configuration", exc_info=True)
             safe_set_attribute(span, "error", str(e))
-            raise HTTPException(status_code=500, detail={"error": str(e)})
+            raise HTTPException(status_code=500, detail={"error": str(e)}) from e
 
 
 @router.get("/clients")
@@ -151,18 +242,11 @@ async def get_mcp_clients():
             # For now, return empty array as expected by frontend
             api_logger.debug("Getting MCP clients - returning empty array")
 
-            return {
-                "clients": [],
-                "total": 0
-            }
+            return {"clients": [], "total": 0}
         except Exception as e:
             api_logger.error(f"Failed to get MCP clients - error={str(e)}")
             safe_set_attribute(span, "error", str(e))
-            return {
-                "clients": [],
-                "total": 0,
-                "error": str(e)
-            }
+            return {"clients": [], "total": 0, "error": str(e)}
 
 
 @router.get("/sessions")
@@ -174,7 +258,7 @@ async def get_mcp_sessions():
 
         try:
             # Basic session info for now
-            status = get_container_status()
+            status = await get_container_status()
 
             session_info = {
                 "active_sessions": 0,  # TODO: Implement real session tracking
@@ -192,7 +276,7 @@ async def get_mcp_sessions():
         except Exception as e:
             api_logger.error(f"Failed to get MCP sessions - error={str(e)}")
             safe_set_attribute(span, "error", str(e))
-            raise HTTPException(status_code=500, detail=str(e))
+            raise HTTPException(status_code=500, detail=str(e)) from e
 
 
 @router.get("/health")

python/src/server/config/config.py

@@ -38,6 +38,23 @@ class RAGStrategyConfig:
     use_reranking: bool = True
 
 
+@dataclass
+class MCPMonitoringConfig:
+    """Configuration for MCP server monitoring strategy.
+
+    Controls how archon-server monitors MCP server status - via HTTP health checks
+    (secure, default) or Docker socket (legacy, security risk).
+
+    Attributes:
+        enable_docker_socket: Whether to use Docker socket for container status.
+                            Default False for security (uses HTTP health checks).
+        health_check_timeout: Timeout in seconds for HTTP health check requests.
+    """
+
+    enable_docker_socket: bool = False
+    health_check_timeout: int = 5
+
+
 def validate_openai_api_key(api_key: str) -> bool:
     """Validate OpenAI API key format."""
     if not api_key:
@@ -68,14 +85,14 @@ def validate_supabase_key(supabase_key: str) -> tuple[bool, str]:
         # Also skip all other validations (aud, exp, etc) since we only care about the role
         decoded = jwt.decode(
             supabase_key,
-            '',
+            "",
             options={
                 "verify_signature": False,
                 "verify_aud": False,
                 "verify_exp": False,
                 "verify_nbf": False,
-                "verify_iat": False
-            }
+                "verify_iat": False,
+            },
         )
         role = decoded.get("role")
 
@@ -232,3 +249,29 @@ def get_rag_strategy_config() -> RAGStrategyConfig:
         use_agentic_rag=str_to_bool(os.getenv("USE_AGENTIC_RAG")),
         use_reranking=str_to_bool(os.getenv("USE_RERANKING")),
     )
+
+
+def get_mcp_monitoring_config() -> MCPMonitoringConfig:
+    """Load MCP monitoring configuration from environment variables.
+
+    Environment Variables:
+        ENABLE_DOCKER_SOCKET_MONITORING: "true"/"false" (default: false)
+            Controls whether to use Docker socket for status monitoring.
+            Default is false for security (uses HTTP health checks instead).
+        MCP_HEALTH_CHECK_TIMEOUT: Timeout in seconds (default: 5)
+            Timeout for HTTP health check requests to MCP server.
+
+    Returns:
+        MCPMonitoringConfig with parsed settings.
+    """
+
+    def str_to_bool(value: str | None) -> bool:
+        """Convert string environment variable to boolean."""
+        if value is None:
+            return False
+        return value.lower() in ("true", "1", "yes", "on")
+
+    return MCPMonitoringConfig(
+        enable_docker_socket=str_to_bool(os.getenv("ENABLE_DOCKER_SOCKET_MONITORING")),
+        health_check_timeout=int(os.getenv("MCP_HEALTH_CHECK_TIMEOUT", "5")),
+    )

python/tests/server/api_routes/test_mcp_api.py

@@ -0,0 +1,381 @@
+"""
+Tests for MCP API endpoints with HTTP and Docker socket modes.
+"""
+
+import os
+import sys
+from unittest.mock import AsyncMock, MagicMock, patch
+
+import httpx
+import pytest
+
+from src.server.api_routes.mcp_api import (
+    get_container_status,
+    get_container_status_docker,
+    get_container_status_http,
+)
+from src.server.config.config import MCPMonitoringConfig
+
+
+@pytest.fixture
+def mock_mcp_url():
+    """Mock MCP URL for testing."""
+    return "http://test-mcp:8051"
+
+
+@pytest.fixture
+def mock_config_http():
+    """Mock configuration with HTTP mode enabled."""
+    return MCPMonitoringConfig(enable_docker_socket=False, health_check_timeout=5)
+
+
+@pytest.fixture
+def mock_config_docker():
+    """Mock configuration with Docker socket mode enabled."""
+    return MCPMonitoringConfig(enable_docker_socket=True, health_check_timeout=5)
+
+
+# HTTP Mode Tests
+
+
+@pytest.mark.asyncio
+async def test_get_container_status_http_running(mock_mcp_url):
+    """Test HTTP health check when MCP server is running."""
+    mock_response = MagicMock()
+    mock_response.json.return_value = {"success": True, "uptime_seconds": 123.45, "health": {}}
+    mock_response.status_code = 200
+
+    with (
+        patch("src.server.api_routes.mcp_api.get_mcp_url", return_value=mock_mcp_url),
+        patch("src.server.api_routes.mcp_api.get_mcp_monitoring_config") as mock_get_config,
+        patch("httpx.AsyncClient") as mock_client_class,
+    ):
+        mock_get_config.return_value = MCPMonitoringConfig(enable_docker_socket=False, health_check_timeout=5)
+
+        # Create mock async context manager
+        mock_client = MagicMock()
+        mock_client.get = AsyncMock(return_value=mock_response)
+        mock_client_class.return_value.__aenter__.return_value = mock_client
+        mock_client_class.return_value.__aexit__.return_value = None
+
+        result = await get_container_status_http()
+
+        assert result["status"] == "running"
+        assert result["uptime"] == 123
+        assert result["logs"] == []
+        mock_client.get.assert_called_once_with(f"{mock_mcp_url}/health")
+
+
+@pytest.mark.asyncio
+async def test_get_container_status_http_unreachable(mock_mcp_url):
+    """Test HTTP health check when MCP server is unreachable."""
+    with (
+        patch("src.server.api_routes.mcp_api.get_mcp_url", return_value=mock_mcp_url),
+        patch("src.server.api_routes.mcp_api.get_mcp_monitoring_config") as mock_get_config,
+        patch("httpx.AsyncClient") as mock_client_class,
+    ):
+        mock_get_config.return_value = MCPMonitoringConfig(enable_docker_socket=False, health_check_timeout=5)
+
+        # Mock connection error
+        mock_client = MagicMock()
+        mock_client.get = AsyncMock(side_effect=httpx.ConnectError("Connection refused"))
+        mock_client_class.return_value.__aenter__.return_value = mock_client
+        mock_client_class.return_value.__aexit__.return_value = None
+
+        result = await get_container_status_http()
+
+        assert result["status"] == "unreachable"
+        assert result["uptime"] is None
+        assert result["logs"] == []
+
+
+@pytest.mark.asyncio
+async def test_get_container_status_http_timeout(mock_mcp_url):
+    """Test HTTP health check when MCP server times out."""
+    with (
+        patch("src.server.api_routes.mcp_api.get_mcp_url", return_value=mock_mcp_url),
+        patch("src.server.api_routes.mcp_api.get_mcp_monitoring_config") as mock_get_config,
+        patch("httpx.AsyncClient") as mock_client_class,
+    ):
+        mock_get_config.return_value = MCPMonitoringConfig(enable_docker_socket=False, health_check_timeout=5)
+
+        # Mock timeout error
+        mock_client = MagicMock()
+        mock_client.get = AsyncMock(side_effect=httpx.TimeoutException("Timeout"))
+        mock_client_class.return_value.__aenter__.return_value = mock_client
+        mock_client_class.return_value.__aexit__.return_value = None
+
+        result = await get_container_status_http()
+
+        assert result["status"] == "unhealthy"
+        assert result["uptime"] is None
+        assert result["logs"] == []
+
+
+@pytest.mark.asyncio
+async def test_get_container_status_http_unhealthy(mock_mcp_url):
+    """Test HTTP health check when MCP server reports unhealthy."""
+    mock_response = MagicMock()
+    mock_response.json.return_value = {"success": False, "error": "Service unavailable"}
+    mock_response.status_code = 200
+
+    with (
+        patch("src.server.api_routes.mcp_api.get_mcp_url", return_value=mock_mcp_url),
+        patch("src.server.api_routes.mcp_api.get_mcp_monitoring_config") as mock_get_config,
+        patch("httpx.AsyncClient") as mock_client_class,
+    ):
+        mock_get_config.return_value = MCPMonitoringConfig(enable_docker_socket=False, health_check_timeout=5)
+
+        mock_client = MagicMock()
+        mock_client.get = AsyncMock(return_value=mock_response)
+        mock_client_class.return_value.__aenter__.return_value = mock_client
+        mock_client_class.return_value.__aexit__.return_value = None
+
+        result = await get_container_status_http()
+
+        assert result["status"] == "unhealthy"
+        assert result["uptime"] is None
+        assert result["logs"] == []
+
+
+@pytest.mark.asyncio
+async def test_get_container_status_http_zero_uptime(mock_mcp_url):
+    """Test HTTP health check preserves 0 uptime for freshly-launched MCP."""
+    mock_response = MagicMock()
+    mock_response.json.return_value = {"success": True, "uptime_seconds": 0, "health": {}}
+    mock_response.status_code = 200
+
+    with (
+        patch("src.server.api_routes.mcp_api.get_mcp_url", return_value=mock_mcp_url),
+        patch("src.server.api_routes.mcp_api.get_mcp_monitoring_config") as mock_get_config,
+        patch("httpx.AsyncClient") as mock_client_class,
+    ):
+        mock_get_config.return_value = MCPMonitoringConfig(enable_docker_socket=False, health_check_timeout=5)
+
+        mock_client = MagicMock()
+        mock_client.get = AsyncMock(return_value=mock_response)
+        mock_client_class.return_value.__aenter__.return_value = mock_client
+        mock_client_class.return_value.__aexit__.return_value = None
+
+        result = await get_container_status_http()
+
+        assert result["status"] == "running"
+        assert result["uptime"] == 0  # Important: 0 should be preserved, not None
+        assert result["logs"] == []
+        mock_client.get.assert_called_once_with(f"{mock_mcp_url}/health")
+
+
+@pytest.mark.asyncio
+async def test_get_container_status_http_error(mock_mcp_url):
+    """Test HTTP health check when an unexpected error occurs."""
+    with (
+        patch("src.server.api_routes.mcp_api.get_mcp_url", return_value=mock_mcp_url),
+        patch("src.server.api_routes.mcp_api.get_mcp_monitoring_config") as mock_get_config,
+        patch("httpx.AsyncClient") as mock_client_class,
+    ):
+        mock_get_config.return_value = MCPMonitoringConfig(enable_docker_socket=False, health_check_timeout=5)
+
+        # Mock unexpected error
+        mock_client = MagicMock()
+        mock_client.get = AsyncMock(side_effect=Exception("Unexpected error"))
+        mock_client_class.return_value.__aenter__.return_value = mock_client
+        mock_client_class.return_value.__aexit__.return_value = None
+
+        result = await get_container_status_http()
+
+        assert result["status"] == "error"
+        assert result["uptime"] is None
+        assert result["logs"] == []
+
+
+# Docker Mode Tests
+
+
+def test_get_container_status_docker_running():
+    """Test Docker socket check when container is running."""
+    mock_container = MagicMock()
+    mock_container.status = "running"
+    mock_container.attrs = {
+        "State": {"StartedAt": "2025-01-01T00:00:00.000000000Z"},
+    }
+
+    mock_docker_client = MagicMock()
+    mock_docker_client.containers.get.return_value = mock_container
+
+    # Create mock docker module with errors submodule
+    mock_docker = MagicMock()
+    mock_docker.from_env.return_value = mock_docker_client
+    mock_docker_errors = MagicMock()
+    mock_docker_errors.NotFound = type("NotFound", (Exception,), {})
+
+    with patch.dict("sys.modules", {"docker": mock_docker, "docker.errors": mock_docker_errors}):
+        result = get_container_status_docker()
+
+        assert result["status"] == "running"
+        assert result["uptime"] is not None  # Uptime should be calculated
+        assert result["logs"] == []
+        mock_docker_client.containers.get.assert_called_once_with("archon-mcp")
+        mock_docker_client.close.assert_called_once()
+
+
+def test_get_container_status_docker_stopped():
+    """Test Docker socket check when container is stopped."""
+    mock_container = MagicMock()
+    mock_container.status = "exited"
+
+    mock_docker_client = MagicMock()
+    mock_docker_client.containers.get.return_value = mock_container
+
+    # Create mock docker module with errors submodule
+    mock_docker = MagicMock()
+    mock_docker.from_env.return_value = mock_docker_client
+    mock_docker_errors = MagicMock()
+    mock_docker_errors.NotFound = type("NotFound", (Exception,), {})
+
+    with patch.dict("sys.modules", {"docker": mock_docker, "docker.errors": mock_docker_errors}):
+        result = get_container_status_docker()
+
+        assert result["status"] == "stopped"
+        assert result["uptime"] is None
+        assert result["logs"] == []
+        mock_docker_client.close.assert_called_once()
+
+
+def test_get_container_status_docker_not_found():
+    """Test Docker socket check when container is not found."""
+    # Create a mock NotFound exception
+    mock_not_found = type("NotFound", (Exception,), {})
+
+    mock_docker_client = MagicMock()
+    mock_docker_client.containers.get.side_effect = mock_not_found("Container not found")
+
+    mock_docker = MagicMock()
+    mock_docker.from_env.return_value = mock_docker_client
+    mock_docker.errors = MagicMock()
+    mock_docker.errors.NotFound = mock_not_found
+
+    with patch.dict("sys.modules", {"docker": mock_docker, "docker.errors": mock_docker.errors}):
+        result = get_container_status_docker()
+
+        assert result["status"] == "not_found"
+        assert result["uptime"] is None
+        assert result["logs"] == []
+        assert "message" in result
+        mock_docker_client.close.assert_called_once()
+
+
+def test_get_container_status_docker_error():
+    """Test Docker socket check when an error occurs."""
+    mock_docker_client = MagicMock()
+    mock_docker_client.containers.get.side_effect = Exception("Docker error")
+
+    # Create mock docker module with errors submodule
+    mock_docker = MagicMock()
+    mock_docker.from_env.return_value = mock_docker_client
+    mock_docker_errors = MagicMock()
+    mock_docker_errors.NotFound = type("NotFound", (Exception,), {})
+
+    with patch.dict("sys.modules", {"docker": mock_docker, "docker.errors": mock_docker_errors}):
+        result = get_container_status_docker()
+
+        assert result["status"] == "error"
+        assert result["uptime"] is None
+        assert result["logs"] == []
+        assert "error" in result
+        mock_docker_client.close.assert_called_once()
+
+
+# Routing Tests
+
+
+@pytest.mark.asyncio
+async def test_get_container_status_routes_to_http(mock_mcp_url):
+    """Test that get_container_status routes to HTTP mode by default."""
+    mock_response = MagicMock()
+    mock_response.json.return_value = {"success": True, "uptime_seconds": 100, "health": {}}
+    mock_response.status_code = 200
+
+    with (
+        patch("src.server.api_routes.mcp_api.get_mcp_url", return_value=mock_mcp_url),
+        patch("src.server.api_routes.mcp_api.get_mcp_monitoring_config") as mock_get_config,
+        patch("httpx.AsyncClient") as mock_client_class,
+    ):
+        mock_get_config.return_value = MCPMonitoringConfig(enable_docker_socket=False, health_check_timeout=5)
+
+        mock_client = MagicMock()
+        mock_client.get = AsyncMock(return_value=mock_response)
+        mock_client_class.return_value.__aenter__.return_value = mock_client
+        mock_client_class.return_value.__aexit__.return_value = None
+
+        result = await get_container_status()
+
+        assert result["status"] == "running"
+        assert result["uptime"] == 100
+        mock_client.get.assert_called_once()
+
+
+@pytest.mark.asyncio
+async def test_get_container_status_routes_to_docker():
+    """Test that get_container_status routes to Docker mode when enabled."""
+    mock_container = MagicMock()
+    mock_container.status = "running"
+    mock_container.attrs = {
+        "State": {"StartedAt": "2025-01-01T00:00:00.000000000Z"},
+    }
+
+    mock_docker_client = MagicMock()
+    mock_docker_client.containers.get.return_value = mock_container
+
+    # Create mock docker module with errors submodule
+    mock_docker = MagicMock()
+    mock_docker.from_env.return_value = mock_docker_client
+    mock_docker_errors = MagicMock()
+    mock_docker_errors.NotFound = type("NotFound", (Exception,), {})
+
+    with (
+        patch("src.server.api_routes.mcp_api.get_mcp_monitoring_config") as mock_get_config,
+        patch.dict("sys.modules", {"docker": mock_docker, "docker.errors": mock_docker_errors}),
+    ):
+        mock_get_config.return_value = MCPMonitoringConfig(enable_docker_socket=True, health_check_timeout=5)
+
+        result = await get_container_status()
+
+        assert result["status"] == "running"
+        mock_docker_client.containers.get.assert_called_once_with("archon-mcp")
+        mock_docker_client.close.assert_called_once()
+
+
+# Environment Variable Tests
+
+
+@pytest.mark.asyncio
+async def test_config_defaults_to_http_mode():
+    """Test that configuration defaults to secure HTTP mode."""
+    # Clear any environment variables
+    os.environ.pop("ENABLE_DOCKER_SOCKET_MONITORING", None)
+    os.environ.pop("MCP_HEALTH_CHECK_TIMEOUT", None)
+
+    from src.server.config.config import get_mcp_monitoring_config
+
+    config = get_mcp_monitoring_config()
+
+    assert config.enable_docker_socket is False
+    assert config.health_check_timeout == 5
+
+
+@pytest.mark.asyncio
+async def test_config_respects_environment_variables():
+    """Test that configuration respects environment variables."""
+    os.environ["ENABLE_DOCKER_SOCKET_MONITORING"] = "true"
+    os.environ["MCP_HEALTH_CHECK_TIMEOUT"] = "10"
+
+    from src.server.config.config import get_mcp_monitoring_config
+
+    config = get_mcp_monitoring_config()
+
+    assert config.enable_docker_socket is True
+    assert config.health_check_timeout == 10
+
+    # Cleanup
+    os.environ.pop("ENABLE_DOCKER_SOCKET_MONITORING", None)
+    os.environ.pop("MCP_HEALTH_CHECK_TIMEOUT", None)

python/uv.lock

@@ -188,7 +188,6 @@ all = [
     { name = "asyncpg" },
     { name = "crawl4ai" },
     { name = "cryptography" },
-    { name = "docker" },
     { name = "factory-boy" },
     { name = "fastapi" },
     { name = "httpx" },
@@ -213,6 +212,7 @@ all = [
     { name = "sse-starlette" },
     { name = "structlog" },
     { name = "supabase" },
+    { name = "tldextract" },
     { name = "uvicorn" },
     { name = "watchfiles" },
 ]
@@ -240,7 +240,6 @@ server = [
     { name = "asyncpg" },
     { name = "crawl4ai" },
     { name = "cryptography" },
-    { name = "docker" },
     { name = "fastapi" },
     { name = "httpx" },
     { name = "logfire" },
@@ -295,7 +294,6 @@ all = [
     { name = "asyncpg", specifier = ">=0.29.0" },
     { name = "crawl4ai", specifier = "==0.7.4" },
     { name = "cryptography", specifier = ">=41.0.0" },
-    { name = "docker", specifier = ">=6.1.0" },
     { name = "factory-boy", specifier = ">=3.3.0" },
     { name = "fastapi", specifier = ">=0.104.0" },
     { name = "httpx", specifier = ">=0.24.0" },
@@ -320,6 +318,7 @@ all = [
     { name = "sse-starlette", specifier = ">=2.3.3" },
     { name = "structlog", specifier = ">=23.1.0" },
     { name = "supabase", specifier = "==2.15.1" },
+    { name = "tldextract", specifier = ">=5.0.0" },
     { name = "uvicorn", specifier = ">=0.24.0" },
     { name = "watchfiles", specifier = ">=0.18" },
 ]
@@ -347,7 +346,6 @@ server = [
     { name = "asyncpg", specifier = ">=0.29.0" },
     { name = "crawl4ai", specifier = "==0.7.4" },
     { name = "cryptography", specifier = ">=41.0.0" },
-    { name = "docker", specifier = ">=6.1.0" },
     { name = "fastapi", specifier = ">=0.104.0" },
     { name = "httpx", specifier = ">=0.24.0" },
     { name = "logfire", specifier = ">=0.30.0" },
@@ -819,20 +817,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/12/b3/231ffd4ab1fc9d679809f356cebee130ac7daa00d6d6f3206dd4fd137e9e/distro-1.9.0-py3-none-any.whl", hash = "sha256:7bffd925d65168f85027d8da9af6bddab658135b840670a223589bc0c8ef02b2", size = 20277 },
 ]
 
-[[package]]
-name = "docker"
-version = "7.1.0"
-source = { registry = "https://pypi.org/simple" }
-dependencies = [
-    { name = "pywin32", marker = "sys_platform == 'win32'" },
-    { name = "requests" },
-    { name = "urllib3" },
-]
-sdist = { url = "https://files.pythonhosted.org/packages/91/9b/4a2ea29aeba62471211598dac5d96825bb49348fa07e906ea930394a83ce/docker-7.1.0.tar.gz", hash = "sha256:ad8c70e6e3f8926cb8a92619b832b4ea5299e2831c14284663184e200546fa6c", size = 117834 }
-wheels = [
-    { url = "https://files.pythonhosted.org/packages/e3/26/57c6fb270950d476074c087527a558ccb6f4436657314bfb6cdf484114c4/docker-7.1.0-py3-none-any.whl", hash = "sha256:c96b93b7f0a746f9e77d325bcfb87422a3d8bd4f03136ae8a85b37f1898d5fc0", size = 147774 },
-]
-
 [[package]]
 name = "ecdsa"
 version = "0.19.1"
@@ -2263,10 +2247,6 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/1e/29/b53a9ca6cd366bfc928823679c6a76c7a4c69f8201c0ba7903ad18ebae2f/pydantic_core-2.41.4-cp314-cp314t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:5729225de81fb65b70fdb1907fcf08c75d498f4a6f15af005aabb1fdadc19dfa", size = 2041183 },
     { url = "https://files.pythonhosted.org/packages/c7/3d/f8c1a371ceebcaf94d6dd2d77c6cf4b1c078e13a5837aee83f760b4f7cfd/pydantic_core-2.41.4-cp314-cp314t-win_amd64.whl", hash = "sha256:de2cfbb09e88f0f795fd90cf955858fc2c691df65b1f21f0aa00b99f3fbc661d", size = 1993542 },
     { url = "https://files.pythonhosted.org/packages/8a/ac/9fc61b4f9d079482a290afe8d206b8f490e9fd32d4fc03ed4fc698214e01/pydantic_core-2.41.4-cp314-cp314t-win_arm64.whl", hash = "sha256:d34f950ae05a83e0ede899c595f312ca976023ea1db100cd5aa188f7005e3ab0", size = 1973897 },
-    { url = "https://files.pythonhosted.org/packages/c4/48/ae937e5a831b7c0dc646b2ef788c27cd003894882415300ed21927c21efa/pydantic_core-2.41.4-graalpy312-graalpy250_312_native-macosx_10_12_x86_64.whl", hash = "sha256:4f5d640aeebb438517150fdeec097739614421900e4a08db4a3ef38898798537", size = 2112087 },
-    { url = "https://files.pythonhosted.org/packages/5e/db/6db8073e3d32dae017da7e0d16a9ecb897d0a4d92e00634916e486097961/pydantic_core-2.41.4-graalpy312-graalpy250_312_native-macosx_11_0_arm64.whl", hash = "sha256:4a9ab037b71927babc6d9e7fc01aea9e66dc2a4a34dff06ef0724a4049629f94", size = 1920387 },
-    { url = "https://files.pythonhosted.org/packages/0d/c1/dd3542d072fcc336030d66834872f0328727e3b8de289c662faa04aa270e/pydantic_core-2.41.4-graalpy312-graalpy250_312_native-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e4dab9484ec605c3016df9ad4fd4f9a390bc5d816a3b10c6550f8424bb80b18c", size = 1951495 },
-    { url = "https://files.pythonhosted.org/packages/2b/c6/db8d13a1f8ab3f1eb08c88bd00fd62d44311e3456d1e85c0e59e0a0376e7/pydantic_core-2.41.4-graalpy312-graalpy250_312_native-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:bd8a5028425820731d8c6c098ab642d7b8b999758e24acae03ed38a66eca8335", size = 2139008 },
 ]
 
 [[package]]