Add security comments for GetResponseAsync/GetStreamingResponseAsync

src/Libraries/Microsoft.Extensions.AI.Abstractions/ChatCompletion/IChatClient.cs

@@ -11,6 +11,11 @@ namespace Microsoft.Extensions.AI;
 /// <summary>Represents a chat client.</summary>
 /// <remarks>
 /// <para>
+/// Applications must consider the risks of prompt injection attacks, data sizes, and the number of messages
+/// sent to the underlying provider or returned from it. Unless a specific <see cref="IChatClient"/> implementation
+/// explicitly documents safeguards for these concerns, the application is expected to implement appropriate protections.
+/// </para>
+/// <para>
 /// Unless otherwise specified, all members of <see cref="IChatClient"/> are thread-safe for concurrent use.
 /// It is expected that all implementations of <see cref="IChatClient"/> support being used by multiple requests concurrently.
 /// Instances must not be disposed of while the instance is still in use.

src/Libraries/Microsoft.Extensions.AI.AzureAIInference/AzureAIInferenceExtensions.cs

@@ -9,6 +9,20 @@ namespace Microsoft.Extensions.AI;
 public static class AzureAIInferenceExtensions
 {
     /// <summary>Gets an <see cref="IChatClient"/> for use with this <see cref="ChatCompletionsClient"/>.</summary>
+    /// <remarks>
+    /// <para>
+    /// No prompt injection defense is performed by default. Applications are expected to implement
+    /// guards against prompt injection attacks as needed.
+    /// </para>
+    /// <para>
+    /// There are no limits or restrictions imposed on data sizes sent to the underlying provider or returned from it
+    /// by default. Applications are expected to apply data size limits as needed.
+    /// </para>
+    /// <para>
+    /// There are no limits or restrictions imposed on the number of messages sent to the underlying provider or returned from
+    /// it by default. Applications are expected to apply rate limits on a per-user or per-session basis as needed.
+    /// </para>
+    /// </remarks>
     /// <param name="chatCompletionsClient">The client.</param>
     /// <param name="modelId">The ID of the model to use. If <see langword="null"/>, it can be provided per request via <see cref="ChatOptions.ModelId"/>.</param>
     /// <returns>An <see cref="IChatClient"/> that can be used to converse via the <see cref="ChatCompletionsClient"/>.</returns>

src/Libraries/Microsoft.Extensions.AI.Ollama/OllamaChatClient.cs

@@ -21,6 +21,20 @@
 namespace Microsoft.Extensions.AI;
 
 /// <summary>Represents an <see cref="IChatClient"/> for Ollama.</summary>
+/// <remarks>
+/// <para>
+/// No prompt injection defense is performed by default. Applications are expected to implement
+/// guards against prompt injection attacks as needed.
+/// </para>
+/// <para>
+/// There are no limits or restrictions imposed on data sizes sent to the underlying provider or returned from it
+/// by default. Applications are expected to apply data size limits as needed.
+/// </para>
+/// <para>
+/// There are no limits or restrictions imposed on the number of messages sent to the underlying provider or returned from
+/// it by default. Applications are expected to apply rate limits on a per-user or per-session basis as needed.
+/// </para>
+/// </remarks>
 public sealed class OllamaChatClient : IChatClient
 {
     private static readonly JsonElement _schemalessJsonResponseFormatValue = JsonDocument.Parse("\"json\"").RootElement;

src/Libraries/Microsoft.Extensions.AI.OpenAI/OpenAIClientExtensions.cs

@@ -14,18 +14,60 @@ namespace Microsoft.Extensions.AI;
 public static class OpenAIClientExtensions
 {
     /// <summary>Gets an <see cref="IChatClient"/> for use with this <see cref="ChatClient"/>.</summary>
+    /// <remarks>
+    /// <para>
+    /// No prompt injection defense is performed by default. Applications are expected to implement
+    /// guards against prompt injection attacks as needed.
+    /// </para>
+    /// <para>
+    /// There are no limits or restrictions imposed on data sizes sent to the underlying provider or returned from it
+    /// by default. Applications are expected to apply data size limits as needed.
+    /// </para>
+    /// <para>
+    /// There are no limits or restrictions imposed on the number of messages sent to the underlying provider or returned from
+    /// it by default. Applications are expected to apply rate limits on a per-user or per-session basis as needed.
+    /// </para>
+    /// </remarks>
     /// <param name="chatClient">The client.</param>
     /// <returns>An <see cref="IChatClient"/> that can be used to converse via the <see cref="ChatClient"/>.</returns>
     public static IChatClient AsIChatClient(this ChatClient chatClient) =>
         new OpenAIChatClient(chatClient);
 
     /// <summary>Gets an <see cref="IChatClient"/> for use with this <see cref="OpenAIResponseClient"/>.</summary>
+    /// <remarks>
+    /// <para>
+    /// No prompt injection defense is performed by default. Applications are expected to implement
+    /// guards against prompt injection attacks as needed.
+    /// </para>
+    /// <para>
+    /// There are no limits or restrictions imposed on data sizes sent to the underlying provider or returned from it
+    /// by default. Applications are expected to apply data size limits as needed.
+    /// </para>
+    /// <para>
+    /// There are no limits or restrictions imposed on the number of messages sent to the underlying provider or returned from
+    /// it by default. Applications are expected to apply rate limits on a per-user or per-session basis as needed.
+    /// </para>
+    /// </remarks>
     /// <param name="responseClient">The client.</param>
     /// <returns>An <see cref="IChatClient"/> that can be used to converse via the <see cref="OpenAIResponseClient"/>.</returns>
     public static IChatClient AsIChatClient(this OpenAIResponseClient responseClient) =>
         new OpenAIResponseChatClient(responseClient);
 
     /// <summary>Gets an <see cref="ISpeechToTextClient"/> for use with this <see cref="AudioClient"/>.</summary>
+    /// <remarks>
+    /// <para>
+    /// No prompt injection defense is performed by default. Applications are expected to implement
+    /// guards against prompt injection attacks as needed.
+    /// </para>
+    /// <para>
+    /// There are no limits or restrictions imposed on data sizes sent to the underlying provider or returned from it
+    /// by default. Applications are expected to apply data size limits as needed.
+    /// </para>
+    /// <para>
+    /// There are no limits or restrictions imposed on the number of messages sent to the underlying provider or returned from
+    /// it by default. Applications are expected to apply rate limits on a per-user or per-session basis as needed.
+    /// </para>
+    /// </remarks>
     /// <param name="audioClient">The client.</param>
     /// <returns>An <see cref="ISpeechToTextClient"/> that can be used to transcribe audio via the <see cref="AudioClient"/>.</returns>
     [Experimental("MEAI001")]