Add diagnostic tools for hook debugging

Added two helper scripts:
- debug_hook.sh: Logs hook calls and connectivity tests
- diagnose.sh: Comprehensive diagnostic tool for troubleshooting

The diagnose script checks:
- Config file existence and content
- Backend connectivity
- Hook installation and permissions
- Manual hook testing
- settings.json configuration
- Debug logs

Usage: bash .claude/hooks/diagnose.sh

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

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

.claude/hooks/debug_hook.sh

@@ -0,0 +1,31 @@
+#!/bin/bash
+# Debug hook to verify hooks are being called
+
+LOG_FILE="/tmp/claude_monitor_debug.log"
+
+# Log timestamp and hook name
+echo "=== $(date) ===" >> "$LOG_FILE"
+echo "Hook called from: ${BASH_SOURCE[0]}" >> "$LOG_FILE"
+
+# Read input
+INPUT=$(cat)
+echo "Input received: $INPUT" >> "$LOG_FILE"
+
+# Check config file
+if [ -f ~/.claude/monitor_url ]; then
+  MONITOR_URL=$(cat ~/.claude/monitor_url)
+  echo "Config file found: $MONITOR_URL" >> "$LOG_FILE"
+else
+  MONITOR_URL="${CLAUDE_MONITOR_URL:-http://localhost:8000}"
+  echo "Using env/default: $MONITOR_URL" >> "$LOG_FILE"
+fi
+
+# Test backend connectivity
+echo "Testing connectivity to: $MONITOR_URL/health" >> "$LOG_FILE"
+RESPONSE=$(curl -s -w "\nHTTP_CODE:%{http_code}" --connect-timeout 2 "$MONITOR_URL/health" 2>&1)
+echo "Response: $RESPONSE" >> "$LOG_FILE"
+
+echo "" >> "$LOG_FILE"
+
+# Pass through to actual hook if this is being tested
+exit 0