claude-mpm 5.6.23__py3-none-any.whl → 5.6.72__py3-none-any.whl

claude_mpm/core/output_style_manager.py

@@ -358,8 +358,11 @@ class OutputStyleManager:
 
             # Only set activeOutputStyle if:
             # 1. No active style is set (first deployment), OR
-            # 2. This is a fresh install (file didn't exist before deployment)
-            should_activate = current_style is None or is_fresh_install
+            # 2. Current style is "default" (not a real user preference), OR
+            # 3. This is a fresh install (file didn't exist before deployment)
+            should_activate = (
+                current_style is None or current_style == "default" or is_fresh_install
+            )
 
             if should_activate and current_style != style_name:
                 settings["activeOutputStyle"] = style_name

claude_mpm/core/socketio_pool.py

@@ -192,9 +192,14 @@ class SocketIOConnectionPool:
         self.health_running = False
         self.last_health_check = datetime.now(timezone.utc)
 
-        # Server configuration
-        self.server_url = None
-        self.server_port = None
+        # Server configuration - use default immediately, update async
+        self.server_port = int(
+            os.environ.get(
+                "CLAUDE_MPM_SOCKETIO_PORT", str(NetworkConfig.DEFAULT_SOCKETIO_PORT)
+            )
+        )
+        self.server_url = f"http://localhost:{self.server_port}"
+        self._port_detection_complete = False
 
         # Pool lifecycle
         self._running = False
@@ -208,7 +213,10 @@ class SocketIOConnectionPool:
             return
 
         self._running = True
-        self._detect_server()
+
+        # Start async port detection in background (non-blocking)
+        # Default port is already set in __init__, this just updates if a better one is found
+        self._detect_server_async()
 
         # Start batch processing thread
         self.batch_running = True
@@ -274,14 +282,29 @@ class SocketIOConnectionPool:
 
         self.logger.info("Socket.IO connection pool stopped")
 
+    def _detect_server_async(self):
+        """Start server detection in background thread.
+
+        This runs port scanning asynchronously to avoid blocking the main thread.
+        The default port is already set in __init__, so this just updates if a better one is found.
+        """
+        threading.Thread(
+            target=self._detect_server, daemon=True, name="port-detect"
+        ).start()
+
     def _detect_server(self):
-        """Detect Socket.IO server configuration."""
-        # Check environment variable first
+        """Detect Socket.IO server configuration.
+
+        This method scans ports to find a running Socket.IO server.
+        It's designed to be run in a background thread to avoid blocking.
+        """
+        # Check environment variable first - if set, use it and skip detection
         env_port = os.environ.get("CLAUDE_MPM_SOCKETIO_PORT")
         if env_port:
             try:
                 self.server_port = int(env_port)
                 self.server_url = f"http://localhost:{self.server_port}"
+                self._port_detection_complete = True
                 self.logger.debug(
                     f"Using Socket.IO server from environment: {self.server_url}"
                 )
@@ -310,19 +333,20 @@ class SocketIOConnectionPool:
         for port in common_ports:
             try:
                 with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s:
-                    s.settimeout(0.05)
+                    # Use 10ms timeout (reduced from 50ms) for faster scanning
+                    s.settimeout(0.01)
                     result = s.connect_ex(("localhost", port))
                     if result == 0:
                         self.server_port = port
                         self.server_url = f"http://localhost:{port}"
+                        self._port_detection_complete = True
                         self.logger.debug(f"Detected Socket.IO server on port {port}")
                         return
             except Exception:  # nosec B112 - intentional: skip ports that fail
                 continue
 
-        # Fall back to default
-        self.server_port = NetworkConfig.DEFAULT_DASHBOARD_PORT
-        self.server_url = f"http://localhost:{self.server_port}"
+        # Keep default port set in __init__, mark detection complete
+        self._port_detection_complete = True
         self.logger.debug(f"Using default Socket.IO server: {self.server_url}")
 
     def _create_client(self) -> Optional[socketio.AsyncClient]:

claude_mpm/hooks/claude_hooks/auto_pause_handler.py

@@ -56,7 +56,7 @@ except ImportError:
 logger = get_logger(__name__)
 
 # Debug mode
-DEBUG = os.environ.get("CLAUDE_MPM_HOOK_DEBUG", "true").lower() != "false"
+DEBUG = os.environ.get("CLAUDE_MPM_HOOK_DEBUG", "false").lower() == "true"
 
 # Warning messages for threshold crossings
 THRESHOLD_WARNINGS = {

claude_mpm/hooks/claude_hooks/event_handlers.py

@@ -3,15 +3,21 @@
 
 This module provides individual event handlers for different types of
 Claude Code hook events.
+
+Supports Dependency Injection:
+- Optional services can be passed via constructor
+- Lazy loading fallback for services not provided
+- Eliminates runtime imports inside methods
 """
 
+import asyncio
 import os
 import re
 import subprocess  # nosec B404 - subprocess used for safe claude CLI version checking only
 import uuid
 from datetime import datetime, timezone
 from pathlib import Path
-from typing import Optional
+from typing import Any, Optional
 
 # Import _log helper to avoid stderr writes (which cause hook errors)
 try:
@@ -42,6 +48,13 @@ except ImportError:
         extract_tool_results,
     )
 
+# Import correlation manager with fallback for direct execution
+# WHY at top level: Runtime relative imports fail with "no known parent package" error
+try:
+    from .correlation_manager import CorrelationManager
+except ImportError:
+    from correlation_manager import CorrelationManager
+
 # Debug mode - MUST match hook_handler.py default (false) to prevent stderr writes
 DEBUG = os.environ.get("CLAUDE_MPM_HOOK_DEBUG", "false").lower() == "true"
 
@@ -54,13 +67,152 @@ except ImportError:
         QUICK_TIMEOUT = 2.0
 
 
+# ============================================================================
+# Optional Dependencies - loaded once at module level for DI
+# ============================================================================
+
+# Log manager (for agent prompt logging)
+_log_manager: Optional[Any] = None
+_log_manager_loaded = False
+
+
+def _get_log_manager() -> Optional[Any]:
+    """Get log manager with lazy loading."""
+    global _log_manager, _log_manager_loaded
+    if not _log_manager_loaded:
+        try:
+            from claude_mpm.core.log_manager import get_log_manager
+
+            _log_manager = get_log_manager()
+        except ImportError:
+            _log_manager = None
+        _log_manager_loaded = True
+    return _log_manager
+
+
+# Config service (for autotodos configuration)
+_config: Optional[Any] = None
+_config_loaded = False
+
+
+def _get_config() -> Optional[Any]:
+    """Get Config with lazy loading."""
+    global _config, _config_loaded
+    if not _config_loaded:
+        try:
+            from claude_mpm.core.config import Config
+
+            _config = Config()
+        except ImportError:
+            _config = None
+        _config_loaded = True
+    return _config
+
+
+# Delegation detector (for anti-pattern detection)
+_delegation_detector: Optional[Any] = None
+_delegation_detector_loaded = False
+
+
+def _get_delegation_detector_service() -> Optional[Any]:
+    """Get delegation detector with lazy loading."""
+    global _delegation_detector, _delegation_detector_loaded
+    if not _delegation_detector_loaded:
+        try:
+            from claude_mpm.services.delegation_detector import get_delegation_detector
+
+            _delegation_detector = get_delegation_detector()
+        except ImportError:
+            _delegation_detector = None
+        _delegation_detector_loaded = True
+    return _delegation_detector
+
+
+# Event log (for PM violation logging)
+_event_log: Optional[Any] = None
+_event_log_loaded = False
+
+
+def _get_event_log_service() -> Optional[Any]:
+    """Get event log with lazy loading."""
+    global _event_log, _event_log_loaded
+    if not _event_log_loaded:
+        try:
+            from claude_mpm.services.event_log import get_event_log
+
+            _event_log = get_event_log()
+        except ImportError:
+            _event_log = None
+        _event_log_loaded = True
+    return _event_log
+
+
 class EventHandlers:
-    """Collection of event handlers for different Claude Code hook events."""
+    """Collection of event handlers for different Claude Code hook events.
 
-    def __init__(self, hook_handler):
-        """Initialize with reference to the main hook handler."""
+    Supports dependency injection for optional services:
+    - log_manager: For agent prompt logging
+    - config: For autotodos configuration
+    - delegation_detector: For anti-pattern detection
+    - event_log: For PM violation logging
+
+    If services are not provided, they are loaded lazily on first use.
+    """
+
+    def __init__(
+        self,
+        hook_handler,
+        *,
+        log_manager: Optional[Any] = None,
+        config: Optional[Any] = None,
+        delegation_detector: Optional[Any] = None,
+        event_log: Optional[Any] = None,
+    ):
+        """Initialize with reference to the main hook handler and optional services.
+
+        Args:
+            hook_handler: The main ClaudeHookHandler instance
+            log_manager: Optional LogManager for agent prompt logging
+            config: Optional Config for autotodos configuration
+            delegation_detector: Optional DelegationDetector for anti-pattern detection
+            event_log: Optional EventLog for PM violation logging
+        """
         self.hook_handler = hook_handler
 
+        # Store injected services (None means use lazy loading)
+        self._log_manager = log_manager
+        self._config = config
+        self._delegation_detector = delegation_detector
+        self._event_log = event_log
+
+    @property
+    def log_manager(self) -> Optional[Any]:
+        """Get log manager (injected or lazy loaded)."""
+        if self._log_manager is not None:
+            return self._log_manager
+        return _get_log_manager()
+
+    @property
+    def config(self) -> Optional[Any]:
+        """Get config (injected or lazy loaded)."""
+        if self._config is not None:
+            return self._config
+        return _get_config()
+
+    @property
+    def delegation_detector(self) -> Optional[Any]:
+        """Get delegation detector (injected or lazy loaded)."""
+        if self._delegation_detector is not None:
+            return self._delegation_detector
+        return _get_delegation_detector_service()
+
+    @property
+    def event_log(self) -> Optional[Any]:
+        """Get event log (injected or lazy loaded)."""
+        if self._event_log is not None:
+            return self._event_log
+        return _get_event_log_service()
+
     def handle_user_prompt_fast(self, event):
         """Handle user prompt with comprehensive data capture.
 
@@ -189,8 +341,6 @@ class EventHandlers:
 
         # Store tool_call_id using CorrelationManager for cross-process retrieval
         if session_id:
-            from .correlation_manager import CorrelationManager
-
             CorrelationManager.store(session_id, tool_call_id, tool_name)
             if DEBUG:
                 _log(
@@ -317,53 +467,50 @@ class EventHandlers:
         )
 
         # Log agent prompt if LogManager is available
-        try:
-            from claude_mpm.core.log_manager import get_log_manager
-
-            log_manager = get_log_manager()
-
-            # Prepare prompt content
-            prompt_content = tool_input.get("prompt", "")
-            if not prompt_content:
-                prompt_content = tool_input.get("description", "")
-
-            if prompt_content:
-                import asyncio
-
-                # Prepare metadata
-                metadata = {
-                    "agent_type": agent_type,
-                    "agent_id": f"{agent_type}_{session_id}",
-                    "session_id": session_id,
-                    "delegation_context": {
-                        "description": tool_input.get("description", ""),
-                        "timestamp": datetime.now(timezone.utc).isoformat(),
-                    },
-                }
+        # Uses injected log_manager or lazy-loaded module-level instance
+        log_manager = self.log_manager
+        if log_manager is not None:
+            try:
+                # Prepare prompt content
+                prompt_content = tool_input.get("prompt", "")
+                if not prompt_content:
+                    prompt_content = tool_input.get("description", "")
+
+                if prompt_content:
+                    # Prepare metadata
+                    metadata = {
+                        "agent_type": agent_type,
+                        "agent_id": f"{agent_type}_{session_id}",
+                        "session_id": session_id,
+                        "delegation_context": {
+                            "description": tool_input.get("description", ""),
+                            "timestamp": datetime.now(timezone.utc).isoformat(),
+                        },
+                    }
 
-                # Log the agent prompt asynchronously
-                try:
-                    loop = asyncio.get_running_loop()
-                    _task = asyncio.create_task(
-                        log_manager.log_prompt(
-                            f"agent_{agent_type}", prompt_content, metadata
-                        )
-                    )  # Fire-and-forget logging (ephemeral hook process)
-                except RuntimeError:
-                    # No running loop, create one
-                    loop = asyncio.new_event_loop()
-                    asyncio.set_event_loop(loop)
-                    loop.run_until_complete(
-                        log_manager.log_prompt(
-                            f"agent_{agent_type}", prompt_content, metadata
+                    # Log the agent prompt asynchronously
+                    try:
+                        loop = asyncio.get_running_loop()
+                        _task = asyncio.create_task(
+                            log_manager.log_prompt(
+                                f"agent_{agent_type}", prompt_content, metadata
+                            )
+                        )  # Fire-and-forget logging (ephemeral hook process)
+                    except RuntimeError:
+                        # No running loop, create one
+                        loop = asyncio.new_event_loop()
+                        asyncio.set_event_loop(loop)
+                        loop.run_until_complete(
+                            log_manager.log_prompt(
+                                f"agent_{agent_type}", prompt_content, metadata
+                            )
                         )
-                    )
 
+                    if DEBUG:
+                        _log(f"  - Agent prompt logged for {agent_type}")
+            except Exception as e:
                 if DEBUG:
-                    _log(f"  - Agent prompt logged for {agent_type}")
-        except Exception as e:
-            if DEBUG:
-                _log(f"  - Could not log agent prompt: {e}")
+                    _log(f"  - Could not log agent prompt: {e}")
 
     def _get_git_branch(self, working_dir: Optional[str] = None) -> str:
         """Get git branch for the given directory with caching."""
@@ -447,8 +594,6 @@ class EventHandlers:
         git_branch = self._get_git_branch(working_dir) if working_dir else "Unknown"
 
         # Retrieve tool_call_id using CorrelationManager for cross-process correlation
-        from .correlation_manager import CorrelationManager
-
         tool_call_id = CorrelationManager.retrieve(session_id) if session_id else None
         if DEBUG and tool_call_id:
             _log(
@@ -600,8 +745,6 @@ class EventHandlers:
                         warning = auto_pause.emit_threshold_warning(threshold_crossed)
                         # CRITICAL: Never write to stderr unconditionally - causes hook errors
                         # Use _log() instead which only writes to file if DEBUG=true
-                        from . import _log
-
                         _log(f"⚠️  Auto-pause threshold crossed: {warning}")
 
                         if DEBUG:
@@ -942,7 +1085,9 @@ class EventHandlers:
         - Provides visibility into new conversation sessions
         - Enables tracking of session lifecycle and duration
         - Useful for monitoring concurrent sessions and resource usage
-        - Auto-inject pending autotodos if enabled in config
+
+        NOTE: This handler is intentionally lightweight - only event monitoring.
+        All initialization/deployment logic runs in MPM CLI startup, not here.
         """
         session_id = event.get("session_id", "")
         working_dir = event.get("cwd", "")
@@ -956,34 +1101,6 @@ class EventHandlers:
             "hook_event_name": "SessionStart",
         }
 
-        # Auto-inject pending autotodos if enabled
-        try:
-            from pathlib import Path
-
-            from claude_mpm.cli.commands.autotodos import get_pending_todos
-            from claude_mpm.core.config import Config
-
-            config = Config()
-            auto_inject_enabled = config.get("autotodos.auto_inject_on_startup", True)
-            max_todos = config.get("autotodos.max_todos_per_session", 10)
-
-            if auto_inject_enabled:
-                # Pass working directory from event to avoid Path.cwd() issues
-                working_dir_param = None
-                if working_dir:
-                    working_dir_param = Path(working_dir)
-
-                pending_todos = get_pending_todos(
-                    max_todos=max_todos, working_dir=working_dir_param
-                )
-                if pending_todos:
-                    session_start_data["pending_autotodos"] = pending_todos
-                    session_start_data["autotodos_count"] = len(pending_todos)
-                    _log(f"  - Auto-injected {len(pending_todos)} pending autotodos")
-        except Exception as e:  # nosec B110
-            # Auto-injection is optional - continue if it fails
-            _log(f"  - Failed to auto-inject autotodos: {e}")
-
         # Debug logging
         _log(f"Hook handler: Processing SessionStart - session: '{session_id}'")
 
@@ -1058,11 +1175,12 @@ class EventHandlers:
         - Delegation patterns = PM doing something WRONG → pm.violation (error)
         - Script failures = Something BROKEN → autotodo.error (todo)
         """
-        # Only scan if delegation detector is available
-        try:
-            from claude_mpm.services.delegation_detector import get_delegation_detector
-            from claude_mpm.services.event_log import get_event_log
-        except ImportError:
+        # Only scan if delegation detector and event log are available
+        # Uses injected services or lazy-loaded module-level instances
+        detector = self.delegation_detector
+        event_log_service = self.event_log
+
+        if detector is None or event_log_service is None:
             if DEBUG:
                 _log("Delegation detector or event log not available")
             return
@@ -1071,22 +1189,16 @@ class EventHandlers:
         if not response_text:
             return
 
-        # Get the delegation detector
-        detector = get_delegation_detector()
-
         # Detect delegation patterns
         detections = detector.detect_user_delegation(response_text)
 
         if not detections:
             return  # No patterns detected
 
-        # Get event log for violation recording
-        event_log = get_event_log()
-
         # Create PM violation events (NOT autotodos)
         for detection in detections:
             # Create event log entry as pm.violation
-            event_log.append_event(
+            event_log_service.append_event(
                 event_type="pm.violation",
                 payload={
                     "violation_type": "delegation_anti_pattern",