@maestro-ai/mcp-server 1.2.0 → 2.0.0

package/dist/content/skills/specialist-dados-analytics-ia/mcp_functions/validate_analytics_quality.py

@@ -0,0 +1,772 @@
+#!/usr/bin/env python3
+"""
+Função MCP de Referência: validate_analytics_quality
+
+Esta função serve como referência para implementação no MCP.
+NÃO EXECUTÁVEL LOCALMENTE - Apenas documentação da estrutura esperada.
+
+Propósito: Validar qualidade do pipeline de analytics
+"""
+
+async def validate_analytics_quality(params: dict) -> dict:
+    """
+    Valida qualidade do pipeline de analytics
+    
+    Args:
+        params: {
+            "pipeline_path": str,                 # Caminho do pipeline
+            "feature_content": dict,               # Conteúdo da feature
+            "validation_level": str,               # basic|complete|strict
+            "data_quality_checks": bool,           # Validar qualidade de dados
+            "pipeline_tests": bool,                 # Testar pipeline
+            "dashboard_validation": bool,           # Validar dashboard
+            "project_context": dict              # Contexto do projeto
+        }
+    
+    Returns:
+        dict: {
+            "success": bool,                      # Status da validação
+            "score": int,                         # Score 0-100
+            "issues": list[dict],                 # Issues encontrados
+            "recommendations": list[str],         # Recomendações
+            "can_proceed": bool,                  # Pode prosseguir
+            "validation_details": dict,           # Detalhes da validação
+            "quality_metrics": dict,              # Métricas de qualidade
+            "compliance_status": dict             # Status de compliance
+        }
+    """
+    
+    # IMPLEMENTAÇÃO ESPERADA NO MCP:
+    
+    feature_content = params["feature_content"]
+    validation_level = params.get("validation_level", "complete")
+    
+    # 1. Validação Estrutural
+    structure_validation = await validate_analytics_structure(feature_content)
+    
+    # 2. Validação de Conteúdo
+    content_validation = await validate_analytics_content(feature_content)
+    
+    # 3. Validação de Lógica
+    logic_validation = await validate_analytics_logic(feature_content)
+    
+    # 4. Validação de Qualidade de Dados (se solicitado)
+    data_quality_validation = {}
+    if params.get("data_quality_checks", True):
+        data_quality_validation = await validate_data_quality(feature_content)
+    
+    # 5. Validação de Pipeline (se solicitado)
+    pipeline_validation = {}
+    if params.get("pipeline_tests", True):
+        pipeline_validation = await validate_pipeline_implementation(feature_content)
+    
+    # 6. Validação de Dashboard (se solicitado)
+    dashboard_validation = {}
+    if params.get("dashboard_validation", True):
+        dashboard_validation = await validate_dashboard_implementation(feature_content)
+    
+    # 7. Calcular score final
+    final_score = calculate_analytics_final_score(
+        structure_validation,
+        content_validation,
+        logic_validation,
+        data_quality_validation,
+        pipeline_validation,
+        dashboard_validation
+    )
+    
+    # 8. Gerar issues
+    issues = generate_analytics_issues(
+        structure_validation,
+        content_validation,
+        logic_validation,
+        data_quality_validation,
+        pipeline_validation,
+        dashboard_validation
+    )
+    
+    # 9. Gerar recomendações
+    recommendations = generate_analytics_recommendations(
+        feature_content,
+        final_score,
+        issues
+    )
+    
+    # 10. Determinar se pode prosseguir
+    can_proceed = determine_analytics_proceed_status(final_score, validation_level, issues)
+    
+    return {
+        "success": True,
+        "score": final_score,
+        "issues": issues,
+        "recommendations": recommendations,
+        "can_proceed": can_proceed,
+        "validation_details": {
+            "structure": structure_validation,
+            "content": content_validation,
+            "logic": logic_validation
+        },
+        "quality_metrics": data_quality_validation,
+        "pipeline_status": pipeline_validation,
+        "dashboard_status": dashboard_validation
+    }
+
+
+# FUNÇÕES DE VALIDAÇÃO
+
+async def validate_analytics_structure(feature_content: dict) -> dict:
+    """
+    Valida estrutura da feature de analytics
+    
+    Args:
+        feature_content: Conteúdo da feature
+        
+    Returns:
+        dict: Resultado da validação estrutural
+    """
+    result = {
+        "valid": True,
+        "score": 0,
+        "errors": [],
+        "warnings": []
+    }
+    
+    # Campos obrigatórios
+    required_fields = [
+        "project_name",
+        "business_objective",
+        "data_sources",
+        "kpi_definitions",
+        "pipeline_config",
+        "data_model"
+    ]
+    
+    for field in required_fields:
+        if not feature_content.get(field):
+            result["errors"].append(f"Campo obrigatório ausente: {field}")
+            result["valid"] = False
+    
+    # Validar tipos
+    type_validations = {
+        "data_sources": list,
+        "kpi_definitions": list,
+        "pipeline_config": dict,
+        "data_model": dict,
+        "dashboard_config": dict
+    }
+    
+    for field, expected_type in type_validations.items():
+        value = feature_content.get(field)
+        if value is not None and not isinstance(value, expected_type):
+            result["errors"].append(f"Tipo inválido em {field}: esperado {expected_type.__name__}")
+            result["valid"] = False
+    
+    # Validar estrutura de fontes de dados
+    data_sources = feature_content.get("data_sources", [])
+    for i, source in enumerate(data_sources):
+        if not isinstance(source, dict):
+            result["errors"].append(f"Fonte de dados {i} deve ser um dicionário")
+            result["valid"] = False
+            continue
+        
+        required_source_fields = ["name", "type", "connection"]
+        for field in required_source_fields:
+            if not source.get(field):
+                result["errors"].append(f"Fonte {i}: campo obrigatório ausente: {field}")
+                result["valid"] = False
+    
+    # Validar estrutura de KPIs
+    kpi_definitions = feature_content.get("kpi_definitions", [])
+    for i, kpi in enumerate(kpi_definitions):
+        if not isinstance(kpi, dict):
+            result["errors"].append(f"KPI {i} deve ser um dicionário")
+            result["valid"] = False
+            continue
+        
+        required_kpi_fields = ["name", "formula", "target", "frequency"]
+        for field in required_kpi_fields:
+            if not kpi.get(field):
+                result["errors"].append(f"KPI {i}: campo obrigatório ausente: {field}")
+                result["valid"] = False
+    
+    # Calcular score estrutural
+    if result["valid"]:
+        result["score"] = 100
+    else:
+        result["score"] = max(0, 100 - (len(result["errors"]) * 10))
+    
+    return result
+
+
+async def validate_analytics_content(feature_content: dict) -> dict:
+    """
+    Valida conteúdo da feature de analytics
+    
+    Args:
+        feature_content: Conteúdo da feature
+        
+    Returns:
+        dict: Resultado da validação de conteúdo
+    """
+    result = {
+        "valid": True,
+        "score": 0,
+        "errors": [],
+        "warnings": []
+    }
+    
+    # Valores permitidos (enums)
+    enum_validations = {
+        "data_sources.type": ["Database", "API", "File", "Stream"],
+        "kpi_definitions.frequency": ["Real-time", "Diário", "Semanal", "Mensal"],
+        "pipeline_config.orchestration": ["Airflow", "Dagster", "Prefect"],
+        "data_model.type": ["Star Schema", "Snowflake", "Data Vault"],
+        "dashboard_config.tool": ["Metabase", "Looker", "Tableau", "Grafana"]
+    }
+    
+    for field, allowed_values in enum_validations.items():
+        value = get_nested_analytics_value(feature_content, field)
+        if value and value not in allowed_values:
+            result["errors"].append(f"Valor inválido em {field}: {value}. Permitidos: {allowed_values}")
+            result["valid"] = False
+    
+    # Validações de formato
+    format_validations = {
+        "project_name": r"^[a-zA-Z0-9\s\-_]+$",
+        "pipeline_config.schedule": r"^(@daily|@hourly|@weekly|\d+\s+\*+\s+\*)$"
+    }
+    
+    import re
+    for field, pattern in format_validations.items():
+        value = feature_content.get(field)
+        if value and not re.match(pattern, value):
+            result["errors"].append(f"Formato inválido em {field}: {value}")
+            result["valid"] = False
+    
+    # Validações de lógica de negócio
+    business_validations = [
+        {
+            "condition": lambda t: len(t.get("data_sources", [])) == 0,
+            "error": "Pelo menos uma fonte de dados deve ser configurada",
+            "field": "data_sources"
+        },
+        {
+            "condition": lambda t: len(t.get("kpi_definitions", [])) == 0,
+            "error": "Pelo menos um KPI deve ser definido",
+            "field": "kpi_definitions"
+        },
+        {
+            "condition": lambda t: t.get("pipeline_config", {}).get("orchestration") and not t.get("pipeline_config", {}).get("schedule"),
+            "error": "Pipeline com orquestração requer schedule definido",
+            "field": "pipeline_config.schedule"
+        }
+    ]
+    
+    for validation in business_validations:
+        if validation["condition"](feature_content):
+            result["errors"].append(validation["error"])
+            result["valid"] = False
+    
+    # Calcular score de conteúdo
+    total_fields = len(enum_validations) + len(format_validations) + len(business_validations)
+    valid_fields = total_fields - len(result["errors"])
+    result["score"] = int((valid_fields / total_fields) * 100) if total_fields > 0 else 0
+    
+    return result
+
+
+async def validate_analytics_logic(feature_content: dict) -> dict:
+    """
+    Valida lógica de negócio da feature de analytics
+    
+    Args:
+        feature_content: Conteúdo da feature
+        
+    Returns:
+        dict: Resultado da validação lógica
+    """
+    result = {
+        "valid": True,
+        "score": 0,
+        "errors": [],
+        "warnings": []
+    }
+    
+    # Regras de consistência
+    consistency_rules = [
+        {
+            "name": "Fontes e KPIs Consistentes",
+            "check": lambda t: (
+                len(t.get("data_sources", [])) > 0 and
+                len(t.get("kpi_definitions", [])) > 0
+            ),
+            "error": "Fontes de dados e KPIs devem ser configurados"
+        },
+        {
+            "name": "Pipeline Completo",
+            "check": lambda t: (
+                t.get("pipeline_config", {}).get("extract") and
+                t.get("pipeline_config", {}).get("transform") and
+                t.get("pipeline_config", {}).get("load")
+            ),
+            "error": "Pipeline deve ter extract, transform e load"
+        },
+        {
+            "name": "Modelo Dimensional Definido",
+            "check": lambda t: (
+                t.get("data_model", {}).get("type") and
+                t.get("data_model", {}).get("fact_tables") and
+                t.get("data_model", {}).get("dimensions")
+            ),
+            "error": "Modelo dimensional deve ter tipo, fatos e dimensões"
+        },
+        {
+            "name": "Dashboard Configurado",
+            "check": lambda t: (
+                t.get("dashboard_config", {}).get("tool") and
+                t.get("dashboard_config", {}).get("kpi_mappings")
+            ),
+            "error": "Dashboard deve ter ferramenta e mapeamento de KPIs"
+        }
+    ]
+    
+    for rule in consistency_rules:
+        if not rule["check"](feature_content):
+            result["errors"].append(f"{rule['name']}: {rule['error']}")
+            result["valid"] = False
+    
+    # Regras de best practices
+    best_practices = [
+        {
+            "name": "Múltiplas Fontes de Dados",
+            "check": lambda t: len(t.get("data_sources", [])) >= 2,
+            "warning": "Recomendado usar múltiplas fontes para analytics robusto"
+        },
+        {
+            "name": "KPIs com Metas",
+            "check": lambda t: all(
+                kpi.get("target") for kpi in t.get("kpi_definitions", [])
+            ),
+            "warning": "Todos os KPIs devem ter metas definidas"
+        },
+        {
+            "name": "Monitoramento Configurado",
+            "check": lambda t: t.get("monitoring_config", {}).get("enabled", False),
+            "warning": "Recomendado habilitar monitoramento do pipeline"
+        },
+        {
+            "name": "Alertas Configurados",
+            "check": lambda t: t.get("alerting_config", {}).get("enabled", False),
+            "warning": "Recomendado configurar alertas para KPIs críticos"
+        }
+    ]
+    
+    for practice in best_practices:
+        if not practice["check"](feature_content):
+            result["warnings"].append(f"{practice['name']}: {practice['warning']}")
+    
+    # Calcular score lógico
+    total_rules = len(consistency_rules) + len(best_practices)
+    passed_rules = sum([
+        rule["check"](feature_content) for rule in consistency_rules
+    ]) + sum([
+        practice["check"](feature_content) for practice in best_practices
+    ])
+    
+    result["score"] = int((passed_rules / total_rules) * 100) if total_rules > 0 else 0
+    
+    return result
+
+
+async def validate_data_quality(feature_content: dict) -> dict:
+    """
+    Valida aspectos de qualidade de dados
+    
+    Args:
+        feature_content: Conteúdo da feature
+        
+    Returns:
+        dict: Resultado da validação de qualidade de dados
+    """
+    result = {
+        "valid": True,
+        "score": 0,
+        "issues": [],
+        "recommendations": []
+    }
+    
+    # Data quality checks
+    quality_checks = [
+        {
+            "name": "Data Completeness",
+            "check": lambda t: t.get("data_quality", {}).get("completeness_threshold", 0) >= 95,
+            "issue": "Threshold de completude muito baixo (< 95%)",
+            "recommendation": "Configure threshold de completude ≥ 95%"
+        },
+        {
+            "name": "Data Accuracy",
+            "check": lambda t: t.get("data_quality", {}).get("accuracy_threshold", 0) >= 99,
+            "issue": "Threshold de acurácia muito baixo (< 99%)",
+            "recommendation": "Configure threshold de acurácia ≥ 99%"
+        },
+        {
+            "name": "Data Freshness",
+            "check": lambda t: t.get("data_quality", {}).get("freshness_hours", 24) <= 1,
+            "issue": "Freshness muito baixo (> 1 hora)",
+            "recommendation": "Configure freshness ≤ 1 hora para dados em tempo real"
+        },
+        {
+            "name": "Data Consistency",
+            "check": lambda t: t.get("data_quality", {}).get("consistency_checks", False),
+            "issue": "Checks de consistência não configurados",
+            "recommendation": "Implemente checks de consistência automatizados"
+        }
+    ]
+    
+    for check in quality_checks:
+        if not check["check"](feature_content):
+            result["issues"].append({
+                "type": check["name"],
+                "description": check["issue"],
+                "recommendation": check["recommendation"],
+                "severity": "high"
+            })
+            result["valid"] = False
+        else:
+            result["recommendations"].append(f"✅ {check['name']} configurado")
+    
+    # Calcular score de qualidade
+    total_checks = len(quality_checks)
+    passed_checks = sum([check["check"](feature_content) for check in quality_checks])
+    result["score"] = int((passed_checks / total_checks) * 100) if total_checks > 0 else 0
+    
+    return result
+
+
+async def validate_pipeline_implementation(feature_content: dict) -> dict:
+    """
+    Valida implementação do pipeline
+    
+    Args:
+        feature_content: Conteúdo da feature
+        
+    Returns:
+        dict: Resultado da validação do pipeline
+    """
+    result = {
+        "valid": True,
+        "score": 0,
+        "metrics": {},
+        "recommendations": []
+    }
+    
+    # Pipeline implementation checks
+    pipeline_checks = [
+        {
+            "name": "Orchestration Tool",
+            "check": lambda t: t.get("pipeline_config", {}).get("orchestration") in ["Airflow", "Dagster", "Prefect"],
+            "recommendation": "Configure ferramenta de orquestração (Airflow/Dagster/Prefect)"
+        },
+        {
+            "name": "Schedule Defined",
+            "check": lambda t: t.get("pipeline_config", {}).get("schedule"),
+            "recommendation": "Defina schedule para execução do pipeline"
+        },
+        {
+            "name": "Error Handling",
+            "check": lambda t: t.get("pipeline_config", {}).get("error_handling", False),
+            "recommendation": "Implemente tratamento de erros robusto"
+        },
+        {
+            "name": "Retry Logic",
+            "check": lambda t: t.get("pipeline_config", {}).get("retry_config", {}).get("enabled", False),
+            "recommendation": "Configure retry logic para falhas temporárias"
+        },
+        {
+            "name": "Monitoring Enabled",
+            "check": lambda t: t.get("pipeline_config", {}).get("monitoring", {}).get("enabled", False),
+            "recommendation": "Habilite monitoramento do pipeline"
+        }
+    ]
+    
+    for check in pipeline_checks:
+        if check["check"](feature_content):
+            result["recommendations"].append(f"✅ {check['name']} configurado")
+        else:
+            result["recommendations"].append(f"⚠️ {check['recommendation']}")
+    
+    # Simular métricas de pipeline
+    result["metrics"] = {
+        "expected_latency": "< 30 minutos",
+        "expected_throughput": "> 1000 registros/segundo",
+        "expected_availability": "99.5%",
+        "expected_recovery_time": "< 5 minutos"
+    }
+    
+    # Calcular score de pipeline
+    total_checks = len(pipeline_checks)
+    configured_checks = sum([check["check"](feature_content) for check in pipeline_checks])
+    result["score"] = int((configured_checks / total_checks) * 100) if total_checks > 0 else 0
+    
+    return result
+
+
+async def validate_dashboard_implementation(feature_content: dict) -> dict:
+    """
+    Valida implementação do dashboard
+    
+    Args:
+        feature_content: Conteúdo da feature
+        
+    Returns:
+        dict: Resultado da validação do dashboard
+    """
+    result = {
+        "valid": True,
+        "score": 0,
+        "metrics": {},
+        "recommendations": []
+    }
+    
+    # Dashboard implementation checks
+    dashboard_checks = [
+        {
+            "name": "BI Tool Selected",
+            "check": lambda t: t.get("dashboard_config", {}).get("tool") in ["Metabase", "Looker", "Tableau", "Grafana"],
+            "recommendation": "Selecione ferramenta de BI (Metabase/Looker/Tableau/Grafana)"
+        },
+        {
+            "name": "KPI Mappings",
+            "check": lambda t: len(t.get("dashboard_config", {}).get("kpi_mappings", [])) > 0,
+            "recommendation": "Mapeie KPIs para visualizações no dashboard"
+        },
+        {
+            "name": "Filters Configured",
+            "check": lambda t: len(t.get("dashboard_config", {}).get("filters", [])) > 0,
+            "recommendation": "Configure filtros interativos no dashboard"
+        },
+        {
+            "name": "Access Control",
+            "check": lambda t: t.get("dashboard_config", {}).get("access_control", {}).get("enabled", False),
+            "recommendation": "Configure controle de acesso para o dashboard"
+        },
+        {
+            "name": "Alerts Configured",
+            "check": lambda t: len(t.get("dashboard_config", {}).get("alerts", [])) > 0,
+            "recommendation": "Configure alertas para KPIs críticos"
+        }
+    ]
+    
+    for check in dashboard_checks:
+        if check["check"](feature_content):
+            result["recommendations"].append(f"✅ {check['name']} configurado")
+        else:
+            result["recommendations"].append(f"⚠️ {check['recommendation']}")
+    
+    # Simular métricas de dashboard
+    result["metrics"] = {
+        "expected_load_time": "< 5 segundos",
+        "expected_concurrent_users": "> 50",
+        "expected_refresh_rate": "Real-time para KPIs críticos",
+        "expected_interactive_response": "< 1 segundo"
+    }
+    
+    # Calcular score de dashboard
+    total_checks = len(dashboard_checks)
+    configured_checks = sum([check["check"](feature_content) for check in dashboard_checks])
+    result["score"] = int((configured_checks / total_checks) * 100) if total_checks > 0 else 0
+    
+    return result
+
+
+# FUNÇÕES AUXILIARES
+
+def get_nested_analytics_value(data: dict, path: str) -> any:
+    """
+    Obtém valor aninhado de dicionário usando path notation
+    
+    Args:
+        data: Dicionário de dados
+        path: Path no formato "level1.level2.level3"
+        
+    Returns:
+        any: Valor encontrado ou None
+    """
+    keys = path.split('.')
+    current = data
+    
+    for key in keys:
+        if isinstance(current, dict) and key in current:
+            current = current[key]
+        else:
+            return None
+    
+    return current
+
+
+def calculate_analytics_final_score(*validations: dict) -> int:
+    """
+    Calcula score final baseado em todas as validações
+    
+    Args:
+        *validations: Resultados das validações
+        
+    Returns:
+        int: Score final 0-100
+    """
+    total_score = 0
+    total_weight = 0
+    
+    weights = {
+        "structure": 20,
+        "content": 25,
+        "logic": 25,
+        "data_quality": 15,
+        "pipeline": 10,
+        "dashboard": 5
+    }
+    
+    validation_types = ["structure", "content", "logic", "data_quality", "pipeline", "dashboard"]
+    
+    for i, validation in enumerate(validations):
+        if validation and "score" in validation:
+            weight = weights.get(validation_types[i], 10)
+            total_score += validation["score"] * weight
+            total_weight += weight
+    
+    return int(total_score / total_weight) if total_weight > 0 else 0
+
+
+def generate_analytics_issues(*validations: dict) -> list[dict]:
+    """
+    Gera lista consolidada de issues
+    
+    Args:
+        *validations: Resultados das validações
+        
+    Returns:
+        list[dict]: Lista de issues
+    """
+    all_issues = []
+    
+    for validation in validations:
+        if validation:
+            if "errors" in validation:
+                for error in validation["errors"]:
+                    all_issues.append({
+                        "type": "error",
+                        "description": error,
+                        "severity": "high"
+                    })
+            
+            if "issues" in validation:
+                all_issues.extend(validation["issues"])
+    
+    return all_issues
+
+
+def generate_analytics_recommendations(feature_content: dict, score: int, issues: list[dict]) -> list[str]:
+    """
+    Gera recomendações baseadas no template e issues
+    
+    Args:
+        feature_content: Conteúdo do template
+        score: Score final
+        issues: Issues encontrados
+        
+    Returns:
+        list[str]: Lista de recomendações
+    """
+    recommendations = []
+    
+    # Recomendações baseadas no score
+    if score < 60:
+        recommendations.append("⚠️ Score baixo (< 60). Reconfiguração obrigatória recomendada.")
+    elif score < 80:
+        recommendations.append("⚠️ Score médio (60-79). Revisar e corrigir issues críticos.")
+    else:
+        recommendations.append("✅ Score bom (≥ 80). Configuração aprovada.")
+    
+    # Recomendações baseadas em issues
+    for issue in issues:
+        if issue["severity"] == "high":
+            recommendations.append(f"🔴 CRÍTICO: {issue['description']}")
+    
+    # Recomendações específicas
+    if not feature_content.get("data_sources"):
+        recommendations.append("🔗 Mapeie todas as fontes de dados disponíveis")
+    
+    if not feature_content.get("kpi_definitions"):
+        recommendations.append("📊 Defina KPIs claros e mensuráveis")
+    
+    if not feature_content.get("pipeline_config"):
+        recommendations.append("🔄 Implemente pipeline ETL/ELT completo")
+    
+    if not feature_content.get("dashboard_config"):
+        recommendations.append("📈 Configure dashboards interativos")
+    
+    return recommendations
+
+
+def determine_analytics_proceed_status(score: int, validation_level: str, issues: list[dict]) -> bool:
+    """
+    Determina se pode prosseguir baseado no score e nível de validação
+    
+    Args:
+        score: Score final
+        validation_level: Nível de validação (basic|complete|strict)
+        issues: Issues encontrados
+        
+    Returns:
+        bool: Pode prosseguir
+    """
+    # Thresholds por nível
+    thresholds = {
+        "basic": 60,
+        "complete": 80,
+        "strict": 90
+    }
+    
+    threshold = thresholds.get(validation_level, 80)
+    
+    # Verificar score mínimo
+    if score < threshold:
+        return False
+    
+    # Verificar issues críticos
+    critical_issues = [issue for issue in issues if issue.get("severity") == "high"]
+    if critical_issues and validation_level == "strict":
+        return False
+    
+    return True
+
+
+# EXEMPLO DE USO
+
+if __name__ == "__main__":
+    """
+    Exemplo de como a função seria chamada no MCP:
+    
+    result = await validate_analytics_quality({
+        "pipeline_path": "analytics/dashboard-analytics/",
+        "feature_content": feature_json,
+        "validation_level": "complete",
+        "data_quality_checks": True,
+        "pipeline_tests": True,
+        "dashboard_validation": True,
+        "project_context": {
+            "industry": "e-commerce",
+            "data_volume": "high",
+            "real_time_requirements": True
+        }
+    })
+    
+    print(f"Validation Success: {result['success']}")
+    print(f"Final Score: {result['score']}")
+    print(f"Can Proceed: {result['can_proceed']}")
+    print(f"Issues Found: {len(result['issues'])}")
+    """
+    
+    print("Esta é uma função de referência para implementação no MCP.")
+    print("NÃO EXECUTÁVEL LOCALMENTE.")
+    print("Consulte MCP_INTEGRATION.md para detalhes de implementação.")