sozsoft-platform/api/modules/Sozsoft.Mcp/Sozsoft.Mcp.Domain/Services/McpSqlGuard.cs

using System;
using System.Collections.Generic;
using System.Text.RegularExpressions;
using Microsoft.Extensions.Localization;
using Sozsoft.Platform.Enums;
using Sozsoft.Platform.Localization;
using Volo.Abp.Domain.Services;

namespace Sozsoft.Mcp.Services;

/// <summary>
/// Salt-okunur sorgu kurali. Hem serbest SELECT hem kayitli nesne cagrisi buradan gecer:
/// ifadenin gercekten okuma oldugu dogrulanir, tanimlayici ve parametre adlari kalibi gecer,
/// satir tavani sorguya yazilir.
/// </summary>
public interface IMcpSqlGuard
{
    /// <summary>Tek bir okuma ifadesi oldugunu dogrular; degilse nedenini dondurur.</summary>
    McpProblem? CheckReadOnlyStatement(string sql);

    /// <summary>Sema/nesne adini dogrular ve saglayiciya gore tirnaklar.</summary>
    McpOutcome<string> QuoteQualifiedName(string schemaName, string objectName, DataSourceTypeEnum dataSourceType);

    /// <summary>Saglayicinin varsayilan semasi.</summary>
    string DefaultSchemaName(DataSourceTypeEnum dataSourceType);

    /// <summary>
    /// Satir tavanini sorgunun kendisine yazar: SQL Server'da <c>TOP (n)</c>, PostgreSQL'de
    /// <c>LIMIT n</c>. Sorgu kendi sinirini zaten tasiyorsa dokunulmaz.
    /// </summary>
    string ApplyRowLimit(string sql, int maxRows, DataSourceTypeEnum dataSourceType);

    /// <summary>
    /// Istemciden gelen parametre adlarini dogrular ve bastaki @ isaretini duserek Dapper'in
    /// bekledigi bicime cevirir.
    /// </summary>
    McpOutcome<Dictionary<string, object>> NormalizeParameters(IReadOnlyDictionary<string, object?>? parameters);
}

public class McpSqlGuard(IStringLocalizer<PlatformResource> localizer) : DomainService, IMcpSqlGuard
{
    /// <summary>Tek parcali SQL tanimlayicisi; sema ve nesne adlari bu kalibi gecmek zorundadir.</summary>
    private static readonly Regex SafeIdentifierRegex = new(
        @"^[A-Za-z_][A-Za-z0-9_]{0,127}$",
        RegexOptions.Compiled | RegexOptions.CultureInvariant);

    private static readonly Regex SafeParameterNameRegex = new(
        @"^@?[A-Za-z_][A-Za-z0-9_]{0,127}$",
        RegexOptions.Compiled | RegexOptions.CultureInvariant);

    private static readonly Regex SqlCommentRegex = new(
        @"--[^\r\n]*|/\*[\s\S]*?\*/",
        RegexOptions.Compiled | RegexOptions.CultureInvariant);

    private static readonly Regex SqlStringLiteralRegex = new(
        @"'(?:[^']|'')*'",
        RegexOptions.Compiled | RegexOptions.CultureInvariant);

    /// <summary>
    /// Okuma sozde kalmasi gereken ifadelerde yasakli anahtar kelimeler. Kontrol yorumlar ve
    /// metin sabitleri cikarildiktan sonra yapilir; aksi halde bir yorum blogu ya da tirnak
    /// icindeki metin yasakli bir kelimeyi gizleyebilir.
    /// </summary>
    private static readonly Regex ForbiddenStatementRegex = new(
        @"\b(INSERT|UPDATE|DELETE|MERGE|UPSERT|CREATE|ALTER|DROP|TRUNCATE|GRANT|REVOKE|DENY|BACKUP|RESTORE|SHUTDOWN|RECONFIGURE|EXEC|EXECUTE|WAITFOR|OPENROWSET|OPENDATASOURCE|OPENQUERY|BULK|INTO)\b",
        RegexOptions.Compiled | RegexOptions.IgnoreCase | RegexOptions.CultureInvariant);

    private static readonly Regex SelectPrefixRegex = new(
        @"^\s*SELECT\s+(?:(?:DISTINCT|ALL)\s+)?",
        RegexOptions.Compiled | RegexOptions.IgnoreCase | RegexOptions.CultureInvariant);

    private static readonly Regex ExistingTopRegex = new(
        @"^\s*SELECT\s+(?:(?:DISTINCT|ALL)\s+)?TOP\s*[(\s]",
        RegexOptions.Compiled | RegexOptions.IgnoreCase | RegexOptions.CultureInvariant);

    private static readonly Regex ExistingLimitRegex = new(
        @"\bLIMIT\s+\d+",
        RegexOptions.Compiled | RegexOptions.IgnoreCase | RegexOptions.CultureInvariant);

    public McpProblem? CheckReadOnlyStatement(string sql)
    {
        if (string.IsNullOrWhiteSpace(sql))
        {
            return McpProblem.From(localizer, McpErrorCodes.EmptyQuery);
        }

        var stripped = SqlStringLiteralRegex.Replace(SqlCommentRegex.Replace(sql, " "), "''");

        if (stripped.Contains(';', StringComparison.Ordinal))
        {
            return McpProblem.From(localizer, McpErrorCodes.SingleStatementOnly);
        }

        var trimmed = stripped.TrimStart();
        var startsWithRead = trimmed.StartsWith("SELECT", StringComparison.OrdinalIgnoreCase)
                          || trimmed.StartsWith("WITH", StringComparison.OrdinalIgnoreCase);

        if (!startsWithRead)
        {
            return McpProblem.From(localizer, McpErrorCodes.ReadOnlyStatementRequired);
        }

        var forbidden = ForbiddenStatementRegex.Match(stripped);

        return forbidden.Success
            ? McpProblem.From(localizer, McpErrorCodes.ForbiddenKeyword, forbidden.Value.ToUpperInvariant())
            : null;
    }

    public McpOutcome<string> QuoteQualifiedName(string schemaName, string objectName, DataSourceTypeEnum dataSourceType)
    {
        foreach (var name in (string?[])[schemaName, objectName])
        {
            if (!SafeIdentifierRegex.IsMatch(name ?? string.Empty))
            {
                return McpProblem.From(localizer, McpErrorCodes.InvalidIdentifier, name ?? string.Empty);
            }
        }

        return dataSourceType == DataSourceTypeEnum.Postgresql
            ? $"\"{schemaName}\".\"{objectName}\""
            : $"[{schemaName}].[{objectName}]";
    }

    public string DefaultSchemaName(DataSourceTypeEnum dataSourceType)
        => dataSourceType == DataSourceTypeEnum.Postgresql ? "public" : "dbo";

    public string ApplyRowLimit(string sql, int maxRows, DataSourceTypeEnum dataSourceType)
    {
        if (dataSourceType == DataSourceTypeEnum.Postgresql)
        {
            return ExistingLimitRegex.IsMatch(sql) ? sql : $"{sql} LIMIT {maxRows}";
        }

        if (ExistingTopRegex.IsMatch(sql))
        {
            return sql;
        }

        var match = SelectPrefixRegex.Match(sql);

        // WITH ile baslayan ya da beklenmeyen bicimdeki sorgularda sarmalama yapilmaz: SQL
        // Server'da ORDER BY tasiyan bir alt sorgu gecersizdir. Sonuc kumesi okunurken yine
        // maxRows satirda kesilir.
        return match.Success
            ? string.Concat(sql.AsSpan(0, match.Length), $"TOP ({maxRows}) ", sql.AsSpan(match.Length))
            : sql;
    }

    public McpOutcome<Dictionary<string, object>> NormalizeParameters(IReadOnlyDictionary<string, object?>? parameters)
    {
        var normalized = new Dictionary<string, object>(StringComparer.OrdinalIgnoreCase);
        if (parameters is null)
        {
            return normalized;
        }

        foreach (var (key, value) in parameters)
        {
            if (key is null || !SafeParameterNameRegex.IsMatch(key))
            {
                return McpProblem.From(localizer, McpErrorCodes.InvalidParameterName, key ?? string.Empty);
            }

            normalized[key.TrimStart('@')] = value!;
        }

        return normalized;
    }
}