csharp-sdk/src/ModelContextProtocol.Core/Protocol/CallToolResult.cs

88 lines
3.9 KiB
C#

using System.Diagnostics.CodeAnalysis;
using System.Text.Json.Nodes;
using System.Text.Json.Serialization;
namespace ModelContextProtocol.Protocol;
/// <summary>
/// Represents the result of a <see cref="RequestMethods.ToolsCall"/> request from a client to invoke a tool provided by the server.
/// </summary>
/// <remarks>
/// <para>
/// Tool execution errors (including input validation errors, API failures, and business logic errors)
/// should be reported inside the result object with <see cref="IsError"/> set to <see langword="true"/>,
/// rather than as a <see cref="JsonRpcError"/>. This allows language models to see error details
/// and potentially self-correct in subsequent requests.
/// </para>
/// <para>
/// To return a validation or business-logic error from a tool method, either throw an <see cref="McpException"/>
/// (whose <see cref="Exception.Message"/> will be included in the error result), or declare the tool's return type
/// as <see cref="CallToolResult"/> so it can be returned directly with <see cref="IsError"/> set to <see langword="true"/>
/// and details in <see cref="Content"/>. Using <see cref="CallToolResult"/> as the return type gives the tool full control
/// over both success and error responses.
/// </para>
/// <para>
/// Protocol-level errors (such as unknown tool names, malformed requests that fail schema validation,
/// or server errors) should be reported as MCP protocol error responses using <see cref="McpErrorCode"/>.
/// </para>
/// <para>
/// See the <see href="https://github.com/modelcontextprotocol/specification/blob/main/schema/">schema</see> for details.
/// </para>
/// </remarks>
public sealed class CallToolResult : Result
{
/// <summary>
/// Gets or sets the response content from the tool call.
/// </summary>
[JsonPropertyName("content")]
public IList<ContentBlock> Content { get; set; } = [];
/// <summary>
/// Gets or sets an optional JSON object representing the structured result of the tool call.
/// </summary>
[JsonPropertyName("structuredContent")]
public JsonNode? StructuredContent { get; set; }
/// <summary>
/// Gets or sets a value that indicates whether the tool call was unsuccessful.
/// </summary>
/// <value>
/// <see langword="true"/> to signify that the tool execution failed; <see langword="false"/> if it was successful.
/// </value>
/// <remarks>
/// <para>
/// Tool execution errors (including input validation errors, API failures, and business logic errors)
/// are reported with this property set to <see langword="true"/> and details in the <see cref="Content"/>
/// property, rather than as protocol-level errors.
/// </para>
/// <para>
/// This design allows language models to receive detailed error feedback and potentially self-correct
/// in subsequent requests. For example, if a date parameter is in the wrong format or out of range,
/// the error message in <see cref="Content"/> can explain the issue, enabling the model to retry
/// with corrected parameters.
/// </para>
/// </remarks>
[JsonPropertyName("isError")]
public bool? IsError { get; set; }
/// <summary>
/// Gets or sets the task data for the newly created task.
/// </summary>
/// <remarks>
/// This property is populated only for task-augmented tool calls. When present, the other properties
/// (<see cref="Content"/>, <see cref="StructuredContent"/>, <see cref="IsError"/>) may not be populated.
/// The actual tool result can be retrieved later via <c>tasks/result</c>.
/// </remarks>
[Experimental(Experimentals.Tasks_DiagnosticId, UrlFormat = Experimentals.Tasks_Url)]
[JsonIgnore]
public McpTask? Task
{
get => TaskCore;
set => TaskCore = value;
}
// See ExperimentalInternalPropertyTests.cs before modifying this property.
[JsonInclude]
[JsonPropertyName("task")]
internal McpTask? TaskCore { get; set; }
}