Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
257 changes: 257 additions & 0 deletions docs/concepts/filters.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,257 @@
---
title: Filters
author: halter73
description: MCP Server Handler Filters
uid: filters
---

# MCP Server Handler Filters

For each handler type in the MCP Server, there are corresponding `AddXXXFilter` methods in `McpServerBuilderExtensions.cs` that allow you to add filters to the handler pipeline. The filters are stored in `McpServerOptions.Filters` and applied during server configuration.

## Available Filter Methods

The following filter methods are available:

- `AddListResourceTemplatesFilter` - Filter for list resource templates handlers
- `AddListToolsFilter` - Filter for list tools handlers
- `AddCallToolFilter` - Filter for call tool handlers
- `AddListPromptsFilter` - Filter for list prompts handlers
- `AddGetPromptFilter` - Filter for get prompt handlers
- `AddListResourcesFilter` - Filter for list resources handlers
- `AddReadResourceFilter` - Filter for read resource handlers
- `AddCompleteFilter` - Filter for completion handlers
- `AddSubscribeToResourcesFilter` - Filter for resource subscription handlers
- `AddUnsubscribeFromResourcesFilter` - Filter for resource unsubscription handlers
- `AddSetLoggingLevelFilter` - Filter for logging level handlers

## Usage

Filters are functions that take a handler and return a new handler, allowing you to wrap the original handler with additional functionality:

```csharp
services.AddMcpServer()
.WithListToolsHandler(async (context, cancellationToken) =>
{
// Your base handler logic
return new ListToolsResult { Tools = GetTools() };
})
.AddListToolsFilter(next => async (context, cancellationToken) =>
{
// Pre-processing logic
Console.WriteLine("Before handler execution");

var result = await next(context, cancellationToken);

// Post-processing logic
Console.WriteLine("After handler execution");
return result;
});
```

## Filter Execution Order

```csharp
services.AddMcpServer()
.WithListToolsHandler(baseHandler)
.AddListToolsFilter(filter1) // Executes first (outermost)
.AddListToolsFilter(filter2) // Executes second
.AddListToolsFilter(filter3); // Executes third (closest to handler)
```

Execution flow: `filter1 -> filter2 -> filter3 -> baseHandler -> filter3 -> filter2 -> filter1`

## Common Use Cases

### Logging
```csharp
.AddListToolsFilter(next => async (context, cancellationToken) =>
{
Console.WriteLine($"Processing request from {context.Meta.ProgressToken}");
var result = await next(context, cancellationToken);
Console.WriteLine($"Returning {result.Tools?.Count ?? 0} tools");
return result;
});
```

### Error Handling
```csharp
.AddCallToolFilter(next => async (context, cancellationToken) =>
{
try
{
return await next(context, cancellationToken);
}
catch (Exception ex)
{
return new CallToolResult
{
Content = new[] { new TextContent { Type = "text", Text = $"Error: {ex.Message}" } },
IsError = true
};
}
});
```

### Performance Monitoring
```csharp
.AddListToolsFilter(next => async (context, cancellationToken) =>
{
var stopwatch = Stopwatch.StartNew();
var result = await next(context, cancellationToken);
stopwatch.Stop();
Console.WriteLine($"Handler took {stopwatch.ElapsedMilliseconds}ms");
return result;
});
```

### Caching
```csharp
.AddListResourcesFilter(next => async (context, cancellationToken) =>
{
var cacheKey = $"resources:{context.Params.Cursor}";
if (cache.TryGetValue(cacheKey, out var cached))
return cached;

var result = await next(context, cancellationToken);
cache.Set(cacheKey, result, TimeSpan.FromMinutes(5));
return result;
});
```

## Built-in Authorization Filters

When using the ASP.NET Core integration (`ModelContextProtocol.AspNetCore`), authorization filters are automatically configured to support `[Authorize]` and `[AllowAnonymous]` attributes on MCP server tools, prompts, and resources.

### Authorization Attributes Support

The MCP server automatically respects the following authorization attributes:

- **`[Authorize]`** - Requires authentication for access
- **`[Authorize(Roles = "RoleName")]`** - Requires specific roles
- **`[Authorize(Policy = "PolicyName")]`** - Requires specific authorization policies
- **`[AllowAnonymous]`** - Explicitly allows anonymous access (overrides `[Authorize]`)

### Tool Authorization

Tools can be decorated with authorization attributes to control access:

```csharp
[McpServerToolType]
public class WeatherTools
{
[McpServerTool, Description("Gets public weather data")]
public static string GetWeather(string location)
{
return $"Weather for {location}: Sunny, 25°C";
}

[McpServerTool, Description("Gets detailed weather forecast")]
[Authorize] // Requires authentication
public static string GetDetailedForecast(string location)
{
return $"Detailed forecast for {location}: ...";
}

[McpServerTool, Description("Manages weather alerts")]
[Authorize(Roles = "Admin")] // Requires Admin role
public static string ManageWeatherAlerts(string alertType)
{
return $"Managing alert: {alertType}";
}
}
```

### Class-Level Authorization

You can apply authorization at the class level, which affects all tools in the class:

```csharp
[McpServerToolType]
[Authorize] // All tools require authentication
public class RestrictedTools
{
[McpServerTool, Description("Restricted tool accessible to authenticated users")]
public static string RestrictedOperation()
{
return "Restricted operation completed";
}

[McpServerTool, Description("Public tool accessible to anonymous users")]
[AllowAnonymous] // Overrides class-level [Authorize]
public static string PublicOperation()
{
return "Public operation completed";
}
}
```

### How Authorization Filters Work

The authorization filters work differently for list operations versus individual operations:

#### List Operations (ListTools, ListPrompts, ListResources)
For list operations, the filters automatically remove unauthorized items from the results. Users only see tools, prompts, or resources they have permission to access.

#### Individual Operations (CallTool, GetPrompt, ReadResource)
For individual operations, the filters return authorization errors when access is denied:

- **Tools**: Returns a `CallToolResult` with `IsError = true` and an error message
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am not sure about returning an authorization error for unlisted primitives. This allows sniffing out the names of primitives that one is not authorized to access. The MCP spec isn't really clear on whether it is allowed to call a tool or fetch a resource not present in the list.

In practice it shouldn't happen, as an LLM should never call a tool that isn't in the list. However many clients cache tool lists, so it probably will happen in cases where the list of available tools can shrink.

From an MCP client host developer perspective I'd rather handle authorization errors at the application layer than the model layer. Especially as many models tend to try to work around errors ("let me try to call the tool in a different way.."). Changing tools to do this would also create symmetry with prompts and resources.

So I think throwing an exception is better than a CallToolResult. This allows the client host application to do step up authorization and similar - which isn't really possible if the error is handled at the model layer.

Copy link
Contributor Author

@halter73 halter73 Aug 26, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Changing tools to do this would also create symmetry with prompts and resources.

Agreed. I almost did that to begin with, but I went with a CallToolResult since it did have an IsError concept unlike prompt and resource results. However, upon further reflection, I like the consistency of throwing, so the client doesn't have to special case tool authz errors vs resource and prompts.

That wouldn't fix the issue you describe about sniffing out names of primitives that require authorization, but that doesn't seem any worse to me than a typical HTTP endpoint where you could sniff for a 401/403 vs a 404.

- **Prompts**: Throws an `McpException` with "Access forbidden" message
- **Resources**: Throws an `McpException` with "Access forbidden" message

### Setup Requirements

To use authorization features, you must configure authentication and authorization in your ASP.NET Core application:

```csharp
var builder = WebApplication.CreateBuilder(args);

builder.Services.AddAuthentication("Bearer")
.AddJwtBearer(options => { /* JWT configuration */ })
.AddMcp(options => { /* Resource metadata configuration */ });
builder.Services.AddAuthorization();

builder.Services.AddMcpServer()
.WithHttpTransport()
.WithTools<WeatherTools>()
.AddCallToolFilter(next => async (context, cancellationToken) =>
{
// Custom call tool logic
return await next(context, cancellationToken);
});
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This codeblock invites the question: "Do I need AddCallToolFilter for [Authorize] to work?"

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The answer is no, but I can remove it.

I think I'm going to change the behavior, so you have to call AddAuthorizationFiliters() for authorization to work rather than just doing it implicitly inside of WithHttpTransport(). However, WithHttpTransport() will add filters in IPostConfigureOptions that will throw if the MatchedPrimitive (or one of the MatchPrimitive instances in the collection) has any relevant metadata and the authorization filters haven't been added.

The benefit is that it way more explicit which filters run before and after auth


var app = builder.Build();

app.MapMcp();
app.Run();
```

### Custom Authorization Filters

You can also create custom authorization filters using the filter methods:

```csharp
.AddCallToolFilter(next => async (context, cancellationToken) =>
{
// Custom authorization logic
if (context.User?.Identity?.IsAuthenticated != true)
{
return new CallToolResult
{
Content = [new TextContent { Text = "Custom: Authentication required" }],
IsError = true
};
}

return await next(context, cancellationToken);
});
```

### RequestContext

Within filters, you have access to:

- `context.User` - The current user's `ClaimsPrincipal`
- `context.Services` - The request's service provider for resolving authorization services
- `context.MatchedPrimitive` - The matched tool/prompt/resource with its metadata including authorization attributes via `context.MatchedPrimitive.Metadata`
4 changes: 4 additions & 0 deletions docs/concepts/toc.yml
Original file line number Diff line number Diff line change
Expand Up @@ -13,3 +13,7 @@ items:
items:
- name: Logging
uid: logging
- name: Server Features
items:
- name: Filters
uid: filters
4 changes: 2 additions & 2 deletions samples/AspNetCoreMcpServer/Properties/launchSettings.json
Original file line number Diff line number Diff line change
Expand Up @@ -7,7 +7,7 @@
"applicationUrl": "http://localhost:3001",
"environmentVariables": {
"ASPNETCORE_ENVIRONMENT": "Development",
"OTEL_SERVICE_NAME": "aspnetcore-mcp-server",
"OTEL_SERVICE_NAME": "aspnetcore-mcp-server"
}
},
"https": {
Expand All @@ -16,7 +16,7 @@
"applicationUrl": "https://localhost:7133;http://localhost:3001",
"environmentVariables": {
"ASPNETCORE_ENVIRONMENT": "Development",
"OTEL_SERVICE_NAME": "aspnetcore-mcp-server",
"OTEL_SERVICE_NAME": "aspnetcore-mcp-server"
}
}
}
Expand Down
Loading
Loading