Compare commits

..
Author SHA1 Message Date
Shawn HenryandGitHub 23c8d97f21 New Microsoft Agent Framework logos 2026-06-07 20:40:13 -07:00
116 changed files with 570 additions and 7737 deletions
Binary file not shown.

Before

Width:  |  Height:  |  Size: 1.5 MiB

After

Width:  |  Height:  |  Size: 136 KiB

+1 -1
View File
@@ -99,7 +99,7 @@
<PackageVersion Include="Microsoft.SemanticKernel.Connectors.InMemory" Version="1.67.0-preview" />
<PackageVersion Include="Microsoft.SemanticKernel.Connectors.Qdrant" Version="1.67.0-preview" />
<!-- Agent SDKs -->
<PackageVersion Include="GitHub.Copilot.SDK" Version="1.0.0" />
<PackageVersion Include="GitHub.Copilot.SDK" Version="1.0.0-beta.2" />
<PackageVersion Include="Microsoft.Agents.CopilotStudio.Client" Version="1.3.171-beta" />
<!-- M365 Agents SDK -->
<PackageVersion Include="AdaptiveCards" Version="3.1.0" />
@@ -6,7 +6,6 @@
<Nullable>enable</Nullable>
<ImplicitUsings>enable</ImplicitUsings>
<NoWarn>$(NoWarn);GHCP001</NoWarn>
</PropertyGroup>
<ItemGroup>
@@ -2,22 +2,21 @@
// This sample shows how to create a GitHub Copilot agent with shell command permissions.
using GitHub.Copilot;
using GitHub.Copilot.Rpc;
using GitHub.Copilot.SDK;
using Microsoft.Agents.AI;
// Permission handler that prompts the user for approval
static Task<PermissionDecision> PromptPermission(PermissionRequest request, PermissionInvocation invocation)
static Task<PermissionRequestResult> PromptPermission(PermissionRequest request, PermissionInvocation invocation)
{
Console.WriteLine($"\n[Permission Request: {request.Kind}]");
Console.Write("Approve? (y/n): ");
string? input = Console.ReadLine()?.Trim().ToUpperInvariant();
PermissionDecision decision = input is "Y" or "YES"
? PermissionDecision.ApproveOnce()
: PermissionDecision.Reject();
PermissionRequestResultKind kind = input is "Y" or "YES"
? PermissionRequestResultKind.Approved
: PermissionRequestResultKind.Rejected;
return Task.FromResult(decision);
return Task.FromResult(new PermissionRequestResult { Kind = kind });
}
// Create and start a Copilot client
@@ -36,7 +36,7 @@ dotnet run
You can customize the agent by providing additional configuration:
```csharp
using GitHub.Copilot;
using GitHub.Copilot.SDK;
using Microsoft.Agents.AI;
// Create and start a Copilot client
@@ -79,10 +79,8 @@ AIAgent agent =
.GetProjectOpenAIClient()
.GetResponsesClient()
.AsIChatClient(deploymentName)
.AsHarnessAgent(new HarnessAgentOptions
.AsHarnessAgent(MaxContextWindowTokens, MaxOutputTokens, new HarnessAgentOptions
{
MaxContextWindowTokens = MaxContextWindowTokens,
MaxOutputTokens = MaxOutputTokens,
Name = "ResearchAgent",
Description = "A research assistant that plans and executes research tasks.",
DisableFileAccess = true, // If enabled, this would allow the agent to read/write files in a working directory
@@ -44,10 +44,8 @@ AIAgent webSearchAgent =
.GetProjectOpenAIClient()
.GetResponsesClient()
.AsIChatClient(deploymentName)
.AsHarnessAgent(new HarnessAgentOptions
.AsHarnessAgent(MaxContextWindowTokens, MaxOutputTokens, new HarnessAgentOptions
{
MaxContextWindowTokens = MaxContextWindowTokens,
MaxOutputTokens = MaxOutputTokens,
Name = "WebSearchAgent",
Description = "An agent that can search the web to find information.",
OpenTelemetrySourceName = TracingSourceName,
@@ -94,10 +92,8 @@ AIAgent parentAgent =
.GetProjectOpenAIClient()
.GetResponsesClient()
.AsIChatClient(deploymentName)
.AsHarnessAgent(new HarnessAgentOptions
.AsHarnessAgent(MaxContextWindowTokens, MaxOutputTokens, new HarnessAgentOptions
{
MaxContextWindowTokens = MaxContextWindowTokens,
MaxOutputTokens = MaxOutputTokens,
Name = "StockPriceResearcher",
Description = "An agent that researches stock prices using background agents.",
OpenTelemetrySourceName = TracingSourceName,
@@ -68,10 +68,8 @@ AIAgent agent =
.GetProjectOpenAIClient()
.GetResponsesClient()
.AsIChatClient(deploymentName)
.AsHarnessAgent(new HarnessAgentOptions
.AsHarnessAgent(MaxContextWindowTokens, MaxOutputTokens, new HarnessAgentOptions
{
MaxContextWindowTokens = MaxContextWindowTokens,
MaxOutputTokens = MaxOutputTokens,
Name = "DataAnalyst",
Description = "A data analyst assistant that reads, analyzes, and processes data files.",
OpenTelemetrySourceName = TracingSourceName,
@@ -89,10 +89,8 @@ AIAgent agent =
.GetProjectOpenAIClient()
.GetResponsesClient()
.AsIChatClient(deploymentName)
.AsHarnessAgent(new HarnessAgentOptions
.AsHarnessAgent(MaxContextWindowTokens, MaxOutputTokens, new HarnessAgentOptions
{
MaxContextWindowTokens = MaxContextWindowTokens,
MaxOutputTokens = MaxOutputTokens,
Name = "CodeExecutionAgent",
Description = "A technical assistant with sandboxed code execution and skill-based workflows.",
OpenTelemetrySourceName = TracingSourceName,
@@ -44,33 +44,18 @@ public static class HostedFoundryMemoryProviderScopes
session => new FoundryMemoryProvider.State(new FoundryMemoryProviderScope(GetRequiredHostedContext(session).ChatId));
/// <summary>
/// Returns a <c>stateInitializer</c> that scopes memories per (user, chat) pair, composing
/// <see cref="HostedSessionContext.UserId"/> and <see cref="HostedSessionContext.ChatId"/> into a
/// single delimiter-safe partition key. Use this when memories should be visible only to the same
/// user within the same conversation.
/// Returns a <c>stateInitializer</c> that scopes memories per (user, chat) pair, using
/// <c>"{UserId}:{ChatId}"</c> as the partition key. Use this when memories should be visible
/// only to the same user within the same conversation.
/// </summary>
/// <remarks>
/// Both identity values are opaque strings that may contain any characters, including the <c>:</c>
/// delimiter. To keep the composite key injective (so two distinct (user, chat) pairs can never
/// collide), each part is escaped (<c>\</c> becomes <c>\\</c>, then <c>:</c> becomes <c>\:</c>) before
/// being joined with a <c>::</c> separator.
/// </remarks>
/// <returns>A delegate suitable for the <c>stateInitializer</c> argument of <see cref="FoundryMemoryProvider"/>.</returns>
public static Func<AgentSession?, FoundryMemoryProvider.State> PerUserAndChat() =>
session =>
{
var ctx = GetRequiredHostedContext(session);
return new FoundryMemoryProvider.State(
new FoundryMemoryProviderScope($"{EscapeScopePart(ctx.UserId)}::{EscapeScopePart(ctx.ChatId)}"));
return new FoundryMemoryProvider.State(new FoundryMemoryProviderScope($"{ctx.UserId}:{ctx.ChatId}"));
};
/// <summary>
/// Escapes special characters in a scope part so that distinct (user, chat) pairs produce distinct
/// composite scope keys. Backslashes are escaped first (<c>\</c> becomes <c>\\</c>), then colons
/// (<c>:</c> becomes <c>\:</c>), ensuring the <c>{user}::{chat}</c> format is unambiguous.
/// </summary>
private static string EscapeScopePart(string part) => part.Replace("\\", "\\\\").Replace(":", "\\:");
private static HostedSessionContext GetRequiredHostedContext(AgentSession? session) =>
session?.GetHostedContext()
?? throw new InvalidOperationException(
@@ -6,7 +6,7 @@ using Microsoft.Agents.AI.GitHub.Copilot;
using Microsoft.Extensions.AI;
using Microsoft.Shared.Diagnostics;
namespace GitHub.Copilot;
namespace GitHub.Copilot.SDK;
/// <summary>
/// Provides extension methods for <see cref="CopilotClient"/>
@@ -9,7 +9,7 @@ using System.Text.Json;
using System.Threading;
using System.Threading.Channels;
using System.Threading.Tasks;
using GitHub.Copilot;
using GitHub.Copilot.SDK;
using Microsoft.Extensions.AI;
using Microsoft.Shared.Diagnostics;
@@ -169,7 +169,7 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
Channel<AgentResponseUpdate> channel = Channel.CreateUnbounded<AgentResponseUpdate>();
// Subscribe to session events
using IDisposable subscription = copilotSession.On<SessionEvent>(evt =>
using IDisposable subscription = copilotSession.On(evt =>
{
switch (evt)
{
@@ -210,7 +210,7 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
string prompt = string.Join("\n", messages.Select(m => m.Text));
// Handle DataContent as attachments
(List<AttachmentFile>? attachments, tempDir) = await ProcessDataContentAttachmentsAsync(
(List<UserMessageAttachmentFile>? attachments, tempDir) = await ProcessDataContentAttachmentsAsync(
messages,
cancellationToken).ConfigureAwait(false);
@@ -262,7 +262,10 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
private async Task EnsureClientStartedAsync(CancellationToken cancellationToken)
{
await this._copilotClient.StartAsync(cancellationToken).ConfigureAwait(false);
if (this._copilotClient.State != ConnectionState.Connected)
{
await this._copilotClient.StartAsync(cancellationToken).ConfigureAwait(false);
}
}
private ResumeSessionConfig CreateResumeConfig()
@@ -272,18 +275,36 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
/// <summary>
/// Copies all supported properties from a source <see cref="SessionConfig"/> into a new instance
/// with <see cref="SessionConfigBase.Streaming"/> set to <c>true</c>.
/// with <see cref="SessionConfig.Streaming"/> set to <c>true</c>.
/// </summary>
internal static SessionConfig CopySessionConfig(SessionConfig source)
{
SessionConfig copy = source.Clone();
copy.Streaming = true;
return copy;
return new SessionConfig
{
Model = source.Model,
ReasoningEffort = source.ReasoningEffort,
Tools = source.Tools,
SystemMessage = source.SystemMessage,
AvailableTools = source.AvailableTools,
ExcludedTools = source.ExcludedTools,
Provider = source.Provider,
OnPermissionRequest = source.OnPermissionRequest,
OnUserInputRequest = source.OnUserInputRequest,
Hooks = source.Hooks,
WorkingDirectory = source.WorkingDirectory,
ConfigDir = source.ConfigDir,
McpServers = source.McpServers,
CustomAgents = source.CustomAgents,
SkillDirectories = source.SkillDirectories,
DisabledSkills = source.DisabledSkills,
InfiniteSessions = source.InfiniteSessions,
Streaming = true
};
}
/// <summary>
/// Copies all supported properties from a source <see cref="SessionConfig"/> into a new
/// <see cref="ResumeSessionConfig"/> with <see cref="SessionConfigBase.Streaming"/> set to <c>true</c>.
/// <see cref="ResumeSessionConfig"/> with <see cref="ResumeSessionConfig.Streaming"/> set to <c>true</c>.
/// </summary>
internal static ResumeSessionConfig CopyResumeSessionConfig(SessionConfig? source)
{
@@ -300,7 +321,7 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
OnUserInputRequest = source?.OnUserInputRequest,
Hooks = source?.Hooks,
WorkingDirectory = source?.WorkingDirectory,
ConfigDirectory = source?.ConfigDirectory,
ConfigDir = source?.ConfigDir,
McpServers = source?.McpServers,
CustomAgents = source?.CustomAgents,
SkillDirectories = source?.SkillDirectories,
@@ -373,10 +394,10 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
AdditionalPropertiesDictionary<long>? additionalCounts = null;
if (usageEvent.Data.CacheWriteTokens is long cacheWriteTokens)
if (usageEvent.Data.CacheWriteTokens is double cacheWriteTokens)
{
additionalCounts ??= [];
additionalCounts[nameof(AssistantUsageData.CacheWriteTokens)] = cacheWriteTokens;
additionalCounts[nameof(AssistantUsageData.CacheWriteTokens)] = (long)cacheWriteTokens;
}
if (usageEvent.Data.Cost is double cost)
@@ -385,10 +406,10 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
additionalCounts[nameof(AssistantUsageData.Cost)] = (long)cost;
}
if (usageEvent.Data.Duration is TimeSpan duration)
if (usageEvent.Data.Duration is double duration)
{
additionalCounts ??= [];
additionalCounts[nameof(AssistantUsageData.Duration)] = (long)duration.TotalMilliseconds;
additionalCounts[nameof(AssistantUsageData.Duration)] = (long)duration;
}
return additionalCounts;
@@ -411,7 +432,7 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
private static SessionConfig? GetSessionConfig(IList<AITool>? tools, string? instructions)
{
List<AIFunctionDeclaration>? mappedTools = tools is { Count: > 0 } ? tools.OfType<AIFunctionDeclaration>().ToList() : null;
List<AIFunction>? mappedTools = tools is { Count: > 0 } ? tools.OfType<AIFunction>().ToList() : null;
SystemMessageConfig? systemMessage = instructions is not null ? new SystemMessageConfig { Mode = SystemMessageMode.Append, Content = instructions } : null;
if (mappedTools is null && systemMessage is null)
@@ -422,11 +443,11 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
return new SessionConfig { Tools = mappedTools, SystemMessage = systemMessage };
}
private static async Task<(List<AttachmentFile>? Attachments, string? TempDir)> ProcessDataContentAttachmentsAsync(
private static async Task<(List<UserMessageAttachmentFile>? Attachments, string? TempDir)> ProcessDataContentAttachmentsAsync(
IEnumerable<ChatMessage> messages,
CancellationToken cancellationToken)
{
List<AttachmentFile>? attachments = null;
List<UserMessageAttachmentFile>? attachments = null;
string? tempDir = null;
foreach (ChatMessage message in messages)
{
@@ -440,7 +461,7 @@ public sealed class GitHubCopilotAgent : AIAgent, IAsyncDisposable
string tempFilePath = await dataContent.SaveToAsync(tempDir, cancellationToken).ConfigureAwait(false);
attachments ??= [];
attachments.Add(new AttachmentFile
attachments.Add(new UserMessageAttachmentFile
{
Path = tempFilePath,
DisplayName = Path.GetFileName(tempFilePath)
@@ -4,7 +4,6 @@
<VersionSuffix>preview</VersionSuffix>
<!-- GitHub.Copilot.SDK only supports .NET 8.0+ -->
<TargetFrameworks>$(TargetFrameworksCore)</TargetFrameworks>
<NoWarn>$(NoWarn);GHCP001</NoWarn>
</PropertyGroup>
<PropertyGroup>
@@ -16,16 +16,23 @@ public static class ChatClientHarnessExtensions
{
/// <summary>
/// Creates a new <see cref="HarnessAgent"/> that wraps this <see cref="IChatClient"/> with a pre-configured
/// pipeline including function invocation, per-service-call chat history persistence, optional in-loop compaction, and a rich set
/// of default context providers and agent decorators.
/// pipeline including function invocation, per-service-call chat history persistence, and in-loop compaction.
/// </summary>
/// <param name="chatClient">
/// The <see cref="IChatClient"/> that provides access to the underlying AI model.
/// </param>
/// <param name="maxContextWindowTokens">
/// The maximum number of tokens the model's context window supports (e.g., 1,050,000 for gpt-5.4).
/// Used to configure the compaction strategy.
/// </param>
/// <param name="maxOutputTokens">
/// The maximum number of output tokens the model can generate per response (e.g., 128,000 for gpt-5.4).
/// Used to configure the compaction strategy.
/// </param>
/// <param name="options">
/// Optional configuration options for the agent, including instructions override, tools,
/// additional context providers, chat history provider, and compaction settings.
/// When <see langword="null"/>, the agent uses built-in default settings with compaction disabled.
/// additional context providers, and chat history provider.
/// When <see langword="null"/>, the agent uses built-in default settings.
/// </param>
/// <param name="loggerFactory">
/// Optional logger factory for creating loggers used by the agent and its components.
@@ -36,8 +43,10 @@ public static class ChatClientHarnessExtensions
/// <returns>A new <see cref="HarnessAgent"/> instance.</returns>
public static HarnessAgent AsHarnessAgent(
this IChatClient chatClient,
int maxContextWindowTokens,
int maxOutputTokens,
HarnessAgentOptions? options = null,
ILoggerFactory? loggerFactory = null,
IServiceProvider? services = null) =>
new(chatClient, options, loggerFactory, services);
new(chatClient, maxContextWindowTokens, maxOutputTokens, options, loggerFactory, services);
}
@@ -18,65 +18,50 @@ namespace Microsoft.Agents.AI;
/// <summary>
/// A pre-configured <see cref="DelegatingAIAgent"/> that wraps a <see cref="ChatClientAgent"/> with
/// function invocation, per-service-call chat history persistence, optional in-loop compaction, and a rich set
/// function invocation, per-service-call chat history persistence, in-loop compaction, and a rich set
/// of default context providers and agent decorators.
/// </summary>
/// <remarks>
/// <para>
/// <see cref="HarnessAgent"/> provides an opinionated, batteries-included agent suitable for
/// interactive agentic scenarios such as research, coding, data analysis, and general task automation.
/// It assembles a full pipeline from a caller-supplied <see cref="IChatClient"/> so that callers
/// only need to configure the parts they want to customize.
/// </para>
/// <para>
/// <strong>Chat client pipeline (inner to outer):</strong>
/// <see cref="HarnessAgent"/> assembles the following pipeline from a caller-supplied <see cref="IChatClient"/>:
/// <list type="number">
/// <item><description><see cref="FunctionInvokingChatClient"/> — automatic function/tool invocation with configurable iteration limits.</description></item>
/// <item><description><see cref="MessageInjectingChatClient"/> — allows external code to inject messages into the conversation mid-stream (e.g., for user interrupts).</description></item>
/// <item><description><see cref="PerServiceCallChatHistoryPersistingChatClient"/> — persists chat history after every individual service call within a function-invocation loop, enabling crash recovery and history inspection.</description></item>
/// <item><description><see cref="AIContextProviderChatClient"/> with a <see cref="CompactionProvider"/> — applies context-window compaction before each call so long function-invocation loops do not overflow the context window. Only included when <see cref="HarnessAgentOptions.MaxContextWindowTokens"/> and <see cref="HarnessAgentOptions.MaxOutputTokens"/> are both provided.</description></item>
/// <item><description><see cref="FunctionInvokingChatClient"/> — automatic function/tool invocation.</description></item>
/// <item><description><see cref="MessageInjectingChatClient"/> — allows external code to inject messages into the conversation mid-stream.</description></item>
/// <item><description><see cref="PerServiceCallChatHistoryPersistingChatClient"/> — persists chat history after every individual service call within a function-invocation loop.</description></item>
/// <item><description><see cref="AIContextProviderChatClient"/> with a <see cref="CompactionProvider"/> — applies context-window compaction before each call so long function-invocation loops do not overflow the context window.</description></item>
/// </list>
/// </para>
/// <para>
/// <strong>Context providers (each enabled by default, individually disableable via <see cref="HarnessAgentOptions"/>):</strong>
/// By default, the following context providers are included (each can be disabled via <see cref="HarnessAgentOptions"/>):
/// <list type="bullet">
/// <item><description><see cref="TodoProvider"/> — persistent todo list that the agent uses to track multi-step plans. Disable with <see cref="HarnessAgentOptions.DisableTodoProvider"/>.</description></item>
/// <item><description><see cref="AgentModeProvider"/> — mode tracking (e.g., "plan" vs "execute") that the agent uses to structure its work. Disable with <see cref="HarnessAgentOptions.DisableAgentModeProvider"/>.</description></item>
/// <item><description><see cref="FileMemoryProvider"/> — file-based session memory allowing the agent to persist notes and artifacts across turns. Disable with <see cref="HarnessAgentOptions.DisableFileMemory"/>.</description></item>
/// <item><description><see cref="FileAccessProvider"/> — shared file access providing read/write tools for a working directory. Disable with <see cref="HarnessAgentOptions.DisableFileAccess"/>.</description></item>
/// <item><description><see cref="AgentSkillsProvider"/> — discovers and loads skill definitions from the file system, enabling dynamic tool sets. Disable with <see cref="HarnessAgentOptions.DisableAgentSkillsProvider"/>.</description></item>
/// <item><description><see cref="TodoProvider"/> — todo list management.</description></item>
/// <item><description><see cref="AgentModeProvider"/> — agent mode tracking (plan/execute).</description></item>
/// <item><description><see cref="FileMemoryProvider"/> — file-based session memory.</description></item>
/// <item><description><see cref="FileAccessProvider"/> — shared file access.</description></item>
/// <item><description><see cref="AgentSkillsProvider"/> — skill discovery and loading.</description></item>
/// </list>
/// </para>
/// <para>
/// <strong>Optional context providers (enabled via <see cref="HarnessAgentOptions"/>):</strong>
/// The agent is also wrapped with the following decorators by default (each can be disabled):
/// <list type="bullet">
/// <item><description><see cref="BackgroundAgentsProvider"/> — enables delegation to background agents for parallel work. Enable by setting <see cref="HarnessAgentOptions.BackgroundAgents"/>.</description></item>
/// <item><description><c>ShellEnvironmentProvider</c> — injects OS/shell/CWD information and a shell execution tool. Enable by setting <c>HarnessAgentOptions.ShellExecutor</c> (.NET only).</description></item>
/// <item><description><see cref="ToolApprovalAgent"/> — "don't ask again" tool approval rules.</description></item>
/// <item><description><see cref="OpenTelemetryAgent"/> — OpenTelemetry instrumentation.</description></item>
/// </list>
/// </para>
/// <para>
/// <strong>Agent decorators (each enabled by default, individually disableable):</strong>
/// <list type="bullet">
/// <item><description><see cref="ToolApprovalAgent"/> — "don't ask again" tool approval rules enabling safe unattended execution. Disable with <see cref="HarnessAgentOptions.DisableToolApproval"/>.</description></item>
/// <item><description><see cref="OpenTelemetryAgent"/> — OpenTelemetry instrumentation following semantic conventions for generative AI. Disable with <see cref="HarnessAgentOptions.DisableOpenTelemetry"/>.</description></item>
/// </list>
/// A <see cref="HostedWebSearchTool"/> is added to the chat options by default (can be disabled via
/// <see cref="HarnessAgentOptions.DisableWebSearch"/>).
/// </para>
/// <para>
/// <strong>Default tools:</strong>
/// <list type="bullet">
/// <item><description><see cref="HostedWebSearchTool"/> — a hosted web search tool added to chat options by default. Disable with <see cref="HarnessAgentOptions.DisableWebSearch"/>.</description></item>
/// </list>
/// The underlying <see cref="ChatClientAgent"/> is configured with
/// <see cref="ChatClientAgentOptions.UseProvidedChatClientAsIs"/> and
/// <see cref="ChatClientAgentOptions.RequirePerServiceCallChatHistoryPersistence"/> set to <see langword="true"/>
/// to match the manually-assembled pipeline.
/// </para>
/// <para>
/// <strong>Chat history:</strong> When no <see cref="HarnessAgentOptions.ChatHistoryProvider"/> is supplied,
/// the agent defaults to an <see cref="InMemoryChatHistoryProvider"/>. If compaction is enabled, the provider
/// is configured with a compaction-based chat reducer to keep in-memory history bounded. Otherwise, no reducer
/// is applied.
/// </para>
/// <para>
/// <strong>Default instructions:</strong> The agent includes built-in system instructions (<see cref="DefaultInstructions"/>)
/// that guide general tool usage and reasoning patterns. These can be overridden via <see cref="HarnessAgentOptions.HarnessInstructions"/>
/// and combined with agent-specific instructions via <see cref="ChatOptions.Instructions"/>.
/// When no <see cref="HarnessAgentOptions.ChatHistoryProvider"/> is supplied, the agent defaults to an
/// <see cref="InMemoryChatHistoryProvider"/> whose chat reducer applies the same compaction strategy,
/// keeping in-memory history from growing unboundedly across sessions.
/// </para>
/// </remarks>
[Experimental(DiagnosticIds.Experiments.AgentsAIExperiments)]
@@ -105,13 +90,21 @@ public sealed class HarnessAgent : DelegatingAIAgent
/// </summary>
/// <param name="chatClient">
/// The <see cref="IChatClient"/> that provides access to the underlying AI model.
/// The agent wraps this client in a function-invocation and per-service-call persistence pipeline.
/// When compaction is enabled via <paramref name="options"/>, a compaction decorator is also added.
/// The agent wraps this client in a function-invocation, per-service-call persistence,
/// and compaction pipeline automatically.
/// </param>
/// <param name="maxContextWindowTokens">
/// The maximum number of tokens the model's context window supports (e.g., 1,050,000 for gpt-5.4).
/// Used to configure the compaction strategy.
/// </param>
/// <param name="maxOutputTokens">
/// The maximum number of output tokens the model can generate per response (e.g., 128,000 for gpt-5.4).
/// Used to configure the compaction strategy and to limit the model's output.
/// </param>
/// <param name="options">
/// Optional configuration options for the agent, including instructions override, tools,
/// additional context providers, chat history provider, and compaction settings.
/// When <see langword="null"/>, the agent uses built-in default settings with compaction disabled.
/// additional context providers, and chat history provider.
/// When <see langword="null"/>, the agent uses built-in default settings.
/// </param>
/// <param name="loggerFactory">
/// Optional logger factory for creating loggers used by the agent and its components.
@@ -123,22 +116,23 @@ public sealed class HarnessAgent : DelegatingAIAgent
/// <paramref name="chatClient"/> is <see langword="null"/>.
/// </exception>
/// <exception cref="ArgumentOutOfRangeException">
/// <see cref="HarnessAgentOptions.MaxContextWindowTokens"/> is not positive, or
/// <see cref="HarnessAgentOptions.MaxOutputTokens"/> is negative or greater than or equal to
/// <see cref="HarnessAgentOptions.MaxContextWindowTokens"/> (when both are provided).
/// <paramref name="maxContextWindowTokens"/> is not positive, or
/// <paramref name="maxOutputTokens"/> is negative or greater than or equal to <paramref name="maxContextWindowTokens"/>.
/// </exception>
public HarnessAgent(IChatClient chatClient, HarnessAgentOptions? options = null, ILoggerFactory? loggerFactory = null, IServiceProvider? services = null)
public HarnessAgent(IChatClient chatClient, int maxContextWindowTokens, int maxOutputTokens, HarnessAgentOptions? options = null, ILoggerFactory? loggerFactory = null, IServiceProvider? services = null)
: base(BuildAgent(
Throw.IfNull(chatClient),
maxContextWindowTokens,
maxOutputTokens,
options,
loggerFactory,
services))
{
}
private static AIAgent BuildAgent(IChatClient chatClient, HarnessAgentOptions? options, ILoggerFactory? loggerFactory, IServiceProvider? services)
private static AIAgent BuildAgent(IChatClient chatClient, int maxContextWindowTokens, int maxOutputTokens, HarnessAgentOptions? options, ILoggerFactory? loggerFactory, IServiceProvider? services)
{
ChatClientAgent innerAgent = BuildInnerAgent(chatClient, options, loggerFactory, services);
ChatClientAgent innerAgent = BuildInnerAgent(chatClient, maxContextWindowTokens, maxOutputTokens, options, loggerFactory, services);
AIAgentBuilder builder = innerAgent.AsBuilder();
@@ -155,35 +149,17 @@ public sealed class HarnessAgent : DelegatingAIAgent
return builder.Build(services);
}
private static ChatClientAgent BuildInnerAgent(IChatClient chatClient, HarnessAgentOptions? options, ILoggerFactory? loggerFactory, IServiceProvider? services)
private static ChatClientAgent BuildInnerAgent(IChatClient chatClient, int maxContextWindowTokens, int maxOutputTokens, HarnessAgentOptions? options, ILoggerFactory? loggerFactory, IServiceProvider? services)
{
// Determine compaction strategy:
// 1. DisableCompaction = true → no compaction
// 2. Custom CompactionStrategy provided → use it (ignore token params)
// 3. Both token params provided → build default ContextWindowCompactionStrategy
// 4. Otherwise → no compaction
CompactionStrategy? compactionStrategy = null;
if (options?.DisableCompaction is not true)
{
if (options?.CompactionStrategy is CompactionStrategy customStrategy)
{
compactionStrategy = customStrategy;
}
else if (options?.MaxContextWindowTokens is int maxCtx && options?.MaxOutputTokens is int maxOut)
{
compactionStrategy = new ContextWindowCompactionStrategy(
maxContextWindowTokens: maxCtx,
maxOutputTokens: maxOut);
}
}
var compactionStrategy = new ContextWindowCompactionStrategy(
maxContextWindowTokens: maxContextWindowTokens,
maxOutputTokens: maxOutputTokens);
ChatHistoryProvider chatHistoryProvider = options?.ChatHistoryProvider
?? (compactionStrategy is not null
? new InMemoryChatHistoryProvider(new InMemoryChatHistoryProviderOptions
{
ChatReducer = compactionStrategy.AsChatReducer(),
})
: new InMemoryChatHistoryProvider());
?? new InMemoryChatHistoryProvider(new InMemoryChatHistoryProviderOptions
{
ChatReducer = compactionStrategy.AsChatReducer(),
});
string harnessInstructions = options?.HarnessInstructions ?? DefaultInstructions;
string? agentInstructions = options?.ChatOptions?.Instructions;
@@ -196,34 +172,20 @@ public sealed class HarnessAgent : DelegatingAIAgent
(false, false) => $"{harnessInstructions}\n\n{agentInstructions}",
};
ChatOptions chatOptions = BuildChatOptions(options, instructions, options?.MaxOutputTokens);
ChatOptions chatOptions = BuildChatOptions(options, instructions, maxOutputTokens);
CompactionProvider? compactionProvider = compactionStrategy is not null
? new CompactionProvider(compactionStrategy, loggerFactory: loggerFactory)
: null;
var compactionProvider = new CompactionProvider(compactionStrategy, loggerFactory: loggerFactory);
IEnumerable<AIContextProvider> contextProviders = BuildContextProviders(options, loggerFactory);
ChatClientBuilder chatClientBuilder = chatClient.AsBuilder();
if (options?.DisableNonApprovalRequiredFunctionBypassing is not true)
{
chatClientBuilder.UseNonApprovalRequiredFunctionBypassing();
}
ChatClientBuilder pipeline = chatClientBuilder
return chatClient
.AsBuilder()
.UseFunctionInvocation(loggerFactory, configure: options?.MaximumIterationsPerRequest is int maxIterations
? ficc => ficc.MaximumIterationsPerRequest = maxIterations
: null)
.UseMessageInjection()
.UsePerServiceCallChatHistoryPersistence();
if (compactionProvider is not null)
{
pipeline = pipeline.UseAIContextProviders(compactionProvider);
}
return pipeline
.UsePerServiceCallChatHistoryPersistence()
.UseAIContextProviders(compactionProvider)
.BuildAIAgent(new ChatClientAgentOptions
{
Id = options?.Id,
@@ -241,15 +203,11 @@ public sealed class HarnessAgent : DelegatingAIAgent
services);
}
private static ChatOptions BuildChatOptions(HarnessAgentOptions? options, string instructions, int? maxOutputTokens)
private static ChatOptions BuildChatOptions(HarnessAgentOptions? options, string instructions, int maxOutputTokens)
{
ChatOptions result = options?.ChatOptions?.Clone() ?? new ChatOptions();
result.Instructions = instructions;
if (maxOutputTokens.HasValue)
{
result.MaxOutputTokens ??= maxOutputTokens.Value;
}
result.MaxOutputTokens ??= maxOutputTokens;
if (options?.DisableWebSearch is not true)
{
@@ -2,7 +2,6 @@
using System.Collections.Generic;
using System.Diagnostics.CodeAnalysis;
using Microsoft.Agents.AI.Compaction;
#if NET
using Microsoft.Agents.AI.Tools.Shell;
#endif
@@ -32,68 +31,6 @@ public sealed class HarnessAgentOptions
/// </summary>
public string? Description { get; set; }
/// <summary>
/// Gets or sets the maximum number of tokens the model's context window supports (e.g., 1,050,000 for gpt-5.4).
/// </summary>
/// <remarks>
/// <para>
/// When both <see cref="MaxContextWindowTokens"/> and <see cref="MaxOutputTokens"/> are provided (and no
/// custom <see cref="CompactionStrategy"/> is set), a default <see cref="ContextWindowCompactionStrategy"/>
/// is constructed from these values to prevent function-invocation loops from overflowing the context window.
/// </para>
/// <para>
/// Ignored when <see cref="CompactionStrategy"/> is provided or when <see cref="DisableCompaction"/> is
/// <see langword="true"/>.
/// </para>
/// </remarks>
public int? MaxContextWindowTokens { get; set; }
/// <summary>
/// Gets or sets the maximum number of output tokens the model can generate per response (e.g., 128,000 for gpt-5.4).
/// </summary>
/// <remarks>
/// <para>
/// When set, this value is used as the default for <see cref="ChatOptions"/>.<see cref="ChatOptions.MaxOutputTokens"/>
/// when not explicitly configured.
/// </para>
/// <para>
/// For compaction purposes, this value is used together with <see cref="MaxContextWindowTokens"/> to construct a
/// default <see cref="ContextWindowCompactionStrategy"/> — but only when no custom <see cref="CompactionStrategy"/>
/// is provided and <see cref="DisableCompaction"/> is <see langword="false"/>.
/// </para>
/// </remarks>
public int? MaxOutputTokens { get; set; }
/// <summary>
/// Gets or sets a custom <see cref="Compaction.CompactionStrategy"/> to use for in-loop context-window compaction.
/// </summary>
/// <remarks>
/// <para>
/// When provided, this strategy is used directly and <see cref="MaxContextWindowTokens"/> and
/// <see cref="MaxOutputTokens"/> are ignored for compaction purposes (<see cref="MaxOutputTokens"/> is still
/// used as the default for <see cref="ChatOptions"/>.<see cref="ChatOptions.MaxOutputTokens"/> if set).
/// </para>
/// <para>
/// When <see langword="null"/> and both <see cref="MaxContextWindowTokens"/> and <see cref="MaxOutputTokens"/>
/// are provided, a default <see cref="ContextWindowCompactionStrategy"/> is constructed from those values.
/// </para>
/// <para>
/// This property is ignored when <see cref="DisableCompaction"/> is <see langword="true"/>.
/// </para>
/// </remarks>
public CompactionStrategy? CompactionStrategy { get; set; }
/// <summary>
/// Gets or sets a value indicating whether in-loop compaction is disabled.
/// </summary>
/// <remarks>
/// When <see langword="true"/>, compaction is disabled regardless of <see cref="CompactionStrategy"/>,
/// <see cref="MaxContextWindowTokens"/>, or <see cref="MaxOutputTokens"/> settings. No
/// <see cref="CompactionProvider"/> is added to the chat client pipeline, and the default
/// <see cref="InMemoryChatHistoryProvider"/> is configured without a chat reducer.
/// </remarks>
public bool DisableCompaction { get; set; }
/// <summary>
/// Gets or sets additional chat options such as tools for the agent to use.
/// </summary>
@@ -131,9 +68,9 @@ public sealed class HarnessAgentOptions
/// Gets or sets the <see cref="ChatHistoryProvider"/> to use for storing chat history.
/// </summary>
/// <remarks>
/// When <see langword="null"/>, the agent defaults to an <see cref="InMemoryChatHistoryProvider"/>.
/// If <see cref="MaxContextWindowTokens"/> and <see cref="MaxOutputTokens"/> are both provided,
/// the default provider is configured with a compaction-based chat reducer; otherwise, no reducer is applied.
/// When <see langword="null"/>, the agent defaults to an <see cref="InMemoryChatHistoryProvider"/>
/// configured with a compaction-based chat reducer derived from the <c>maxContextWindowTokens</c>
/// and <c>maxOutputTokens</c> constructor parameters of <see cref="HarnessAgent"/>.
/// </remarks>
public ChatHistoryProvider? ChatHistoryProvider { get; set; }
@@ -173,20 +110,6 @@ public sealed class HarnessAgentOptions
/// </remarks>
public ToolApprovalAgentOptions? ToolApprovalAgentOptions { get; set; }
/// <summary>
/// Gets or sets a value indicating whether bypassing of approval requests for tools that do not
/// require approval is disabled.
/// </summary>
/// <remarks>
/// When <see langword="false"/> (the default), the underlying chat client pipeline includes the decorator
/// added by <see cref="ChatClientBuilderExtensions.UseNonApprovalRequiredFunctionBypassing"/> above the
/// function invocation middleware.
/// This stores automatically approved function calls for tools that do not require approval in the session
/// state when they are returned alongside tools that do, so that only tools that truly require human
/// approval are surfaced to the caller.
/// </remarks>
public bool DisableNonApprovalRequiredFunctionBypassing { get; set; }
/// <summary>
/// Gets or sets a value indicating whether the <see cref="FileMemoryProvider"/> is disabled.
/// </summary>
@@ -21,18 +21,6 @@ namespace Microsoft.Agents.AI.Hosting;
/// from the ambient <see cref="HttpContext"/>.
/// </para>
/// <para>
/// <strong>Security warning:</strong> The configured <see cref="ClaimsIdentitySessionIsolationKeyProviderOptions.ClaimType"/>
/// must uniquely identify the principal within the served population. Display names, usernames, email
/// aliases, and other mutable or non-unique claims are <strong>unsafe</strong> isolation keys unless the
/// host can prove their uniqueness across all callers: two distinct principals that share the same value
/// would receive the same isolation key and could read or overwrite one another's persisted sessions.
/// The default claim type is <see cref="ClaimTypes.NameIdentifier"/>, a stable unique subject identifier
/// that is typically populated from the OpenID Connect <c>sub</c> claim via the default JWT inbound claim
/// mapping (note that this differs from Entra's object identifier <c>oid</c> claim; override
/// <see cref="ClaimsIdentitySessionIsolationKeyProviderOptions.ClaimType"/> if you need <c>oid</c> or your
/// provider maps a different claim).
/// </para>
/// <para>
/// If the <see cref="HttpContext"/> is unavailable, the user is not authenticated, or the specified claim
/// is missing, the provider returns <see langword="null"/>. The consuming <see cref="IsolationKeyScopedAgentSessionStore"/>
/// will then enforce strict or pass-through behavior based on its configuration.
@@ -72,24 +60,18 @@ public class ClaimsIdentitySessionIsolationKeyProvider : SessionIsolationKeyProv
/// <param name="cancellationToken">The <see cref="CancellationToken"/> to monitor for cancellation requests.</param>
/// <returns>
/// A task that represents the asynchronous operation. The task result contains the value of the
/// configured claim type from the current user's identity, or <see langword="null"/> if the HTTP
/// context is unavailable, the user is not authenticated, or the claim is not present.
/// configured claim type from the current user's identity, or <see langword="null"/> if the claim
/// is not present or the HTTP context is unavailable.
/// </returns>
/// <remarks>
/// This method only reads claims from an authenticated principal: if the current request has no
/// authenticated user, it returns <see langword="null"/> rather than trusting claims on an
/// unauthenticated identity. The claim value is retrieved from <c>HttpContext.User.Claims</c>; if
/// multiple claims of the specified type exist, the first match is returned.
/// This method retrieves the claim value from <c>HttpContext.User.Claims</c>. If multiple claims
/// of the specified type exist, the first match is returned.
/// </remarks>
public override ValueTask<string?> GetSessionIsolationKeyAsync(CancellationToken cancellationToken = default)
{
ClaimsPrincipal? user = this._httpContextAccessor?.HttpContext?.User;
if (user?.Identity?.IsAuthenticated != true)
{
return new ValueTask<string?>((string?)null);
}
Claim? claim = user?.Claims.FirstOrDefault(c => c.Type == this._claimType);
Claim? claim = this._httpContextAccessor?
.HttpContext?
.User?.Claims.FirstOrDefault(c => c.Type == this._claimType);
return new ValueTask<string?>(claim?.Value);
}
@@ -14,30 +14,17 @@ public class ClaimsIdentitySessionIsolationKeyProviderOptions
/// </summary>
/// <remarks>
/// <para>
/// Defaults to <see cref="ClaimTypes.NameIdentifier"/>, which corresponds to a stable, unique
/// subject identifier for the authenticated principal. For OpenID Connect tokens (including those
/// issued by Microsoft Entra ID), this is typically populated from the <c>sub</c> claim via the
/// default JWT inbound claim mapping. Note that <c>sub</c> is distinct from Entra's object
/// identifier (<c>oid</c>) claim; if you require the <c>oid</c> claim, or your provider does not map
/// a unique identifier onto <see cref="ClaimTypes.NameIdentifier"/>, override <see cref="ClaimType"/>
/// with the appropriate claim type.
/// </para>
/// <para>
/// <strong>Security warning:</strong> The configured claim must uniquely identify the principal
/// within the served population. Display names (<see cref="ClaimsIdentity.DefaultNameClaimType"/>
/// / <see cref="ClaimTypes.Name"/>), usernames, email aliases, and other mutable or non-unique
/// claims are <strong>unsafe</strong> isolation keys unless the host can prove their uniqueness
/// across all callers. Two distinct principals that share the same value for a non-unique claim
/// would receive the same session-isolation key and could read or overwrite one another's
/// persisted sessions. Only override this value with a claim that is guaranteed unique and stable.
/// Defaults to <see cref="ClaimsIdentity.DefaultNameClaimType"/>, which typically corresponds to
/// the user's name or unique identifier claim.
/// </para>
/// <para>
/// Common alternatives include:
/// <list type="bullet">
/// <item><description>A composite of tenant and subject identifiers — required for multi-tenant hosts where the subject is only unique per tenant</description></item>
/// <item><description>Custom claim types specific to your authentication provider, provided they are unique and stable</description></item>
/// <item><description><c>ClaimTypes.NameIdentifier</c> — Stable user identifier</description></item>
/// <item><description><c>ClaimTypes.Email</c> — Email address</description></item>
/// <item><description>Custom claim types specific to your authentication provider</description></item>
/// </list>
/// </para>
/// </remarks>
public string ClaimType { get; set; } = ClaimTypes.NameIdentifier;
public string ClaimType { get; set; } = ClaimsIdentity.DefaultNameClaimType;
}
@@ -1,7 +1,6 @@
// Copyright (c) Microsoft. All rights reserved.
using System;
using System.Security.Claims;
using Microsoft.AspNetCore.Http;
using Microsoft.Extensions.DependencyInjection;
@@ -20,28 +19,8 @@ public static class ServiceCollectionExtensions
/// <param name="options"> Optional configuration for the claims-based session isolation key provider.</param>
/// <returns>The <see cref="IServiceCollection"/> so that additional calls can be chained.</returns>
/// <remarks>
/// <para>
/// This method requires <see cref="IHttpContextAccessor"/> to be registered in the service collection.
/// Ensure that <c>services.AddHttpContextAccessor()</c> has been called before using this method.
/// </para>
/// <para>
/// When <paramref name="options"/> is not supplied, the isolation key is derived from the
/// <see cref="ClaimTypes.NameIdentifier"/> claim, a stable unique subject identifier. For OpenID
/// Connect tokens (including Microsoft Entra ID), this is typically mapped from the <c>sub</c> claim
/// by the default JWT inbound claim mapping. Authentication schemes that do not project a unique
/// identifier onto <see cref="ClaimTypes.NameIdentifier"/> (or hosts that require a different claim
/// such as Entra's <c>oid</c>) should override
/// <see cref="ClaimsIdentitySessionIsolationKeyProviderOptions.ClaimType"/>; otherwise the key may be
/// absent, which causes strict-mode session stores to fail.
/// </para>
/// <para>
/// <strong>Security warning:</strong> If you override
/// <see cref="ClaimsIdentitySessionIsolationKeyProviderOptions.ClaimType"/>, the chosen claim must
/// uniquely identify the principal within the served population. Display names, usernames, email
/// aliases, and other mutable or non-unique claims are <strong>unsafe</strong> isolation keys unless
/// the host can prove their uniqueness across all callers, because distinct principals that share the
/// same claim value would receive the same isolation key and could access one another's sessions.
/// </para>
/// </remarks>
public static IServiceCollection UseClaimsBasedSessionIsolation(
this IServiceCollection services,
@@ -49,7 +49,7 @@ internal sealed class ForeachExecutor : DeclarativeActionExecutor<Foreach>
EvaluationResult<DataValue> expressionResult = this.Evaluator.GetValue(this.Model.Items);
if (expressionResult.Value is TableDataValue tableValue)
{
this._values = [.. tableValue.Values.Select(ToLoopValue)];
this._values = [.. tableValue.Values.Select(value => value.ToFormula())];
}
else
{
@@ -99,15 +99,6 @@ internal sealed class ForeachExecutor : DeclarativeActionExecutor<Foreach>
}
}
// Power Fx wraps scalar array literals (`=[1, 2, 3]`) as `Table({Value: 1}, ...)`. Unwrap that single-column
// `Value`-record shape so `Local.LoopValue` is the scalar; multi-field and other shapes pass through unchanged.
private static FormulaValue ToLoopValue(DataValue value) =>
value is RecordDataValue record
&& record.Properties.Count == 1
&& record.Properties.TryGetValue("Value", out DataValue? singleColumn)
? singleColumn.ToFormula()
: value.ToFormula();
/// <inheritdoc/>
/// <remarks>
/// Persists the iteration cursor (<see cref="_index"/>), the materialized item snapshot
@@ -5,5 +5,4 @@ namespace Microsoft.Agents.AI.Workflows.Specialized.Magentic;
internal static class MagenticConstants
{
public const string MagenticTaskContextKey = nameof(MagenticTaskContextKey);
public const string CurrentSpeakerStateKey = nameof(CurrentSpeakerStateKey);
}
@@ -90,7 +90,6 @@ internal class MagenticOrchestrator(AIAgent managerAgent, List<AIAgent> team, Ta
private MagenticTaskContext? _taskContext;
private PortBinding? _planReviewPort;
private string? _currentSpeakerExecutorId;
protected override ProtocolBuilder ConfigureProtocol(ProtocolBuilder protocolBuilder)
{
@@ -197,46 +196,15 @@ internal class MagenticOrchestrator(AIAgent managerAgent, List<AIAgent> team, Ta
else
{
// Subsequent turns: agent returned control, go directly to coordination (progress ledger only, no replan).
// Capture the participant's reply into the manager-visible chat history so the progress ledger can see it.
if (messages is { Count: > 0 })
{
// Capture the participant's reply into the manager-visible chat history so the progress ledger can see it.
this._taskContext.ChatHistory.AddRange(messages);
// Share the reply with the other participants except the replier
await this.BroadcastReplyToOtherParticipantsAsync(messages, context, cancellationToken).ConfigureAwait(false);
}
await this.RunCoordinationRoundAsync(this._taskContext, context, cancellationToken).ConfigureAwait(false);
}
}
/// <summary>
/// Forwards a participant's reply to every other participant so they share the running conversation.
/// The messages are buffered (no <see cref="TurnToken"/> is sent) - they only become context for the participant's next turn.
/// </summary>
private ValueTask BroadcastReplyToOtherParticipantsAsync(
List<ChatMessage> messages, IWorkflowContext context, CancellationToken cancellationToken)
{
// Without a known current speaker we cannot exclude the reply's author, so skip the broadcast
// rather than risk echoing the reply back to its own author. This covers the window after a
// checkpoint restore but before any delegation has set the current speaker.
if (string.IsNullOrEmpty(this._currentSpeakerExecutorId))
{
return default;
}
List<Task>? sendTasks = null;
foreach (AIAgent agent in team)
{
string executorId = AIAgentHostExecutor.IdFor(agent);
if (string.Equals(executorId, this._currentSpeakerExecutorId, StringComparison.Ordinal))
{
continue;
}
(sendTasks ??= []).Add(context.SendMessageAsync(messages, executorId, cancellationToken).AsTask());
}
return sendTasks is null ? default : new ValueTask(Task.WhenAll(sendTasks));
}
private ChatMessage? _fullTaskLedgerMessage;
private ValueTask DelegateToTeamAsync(MagenticTaskContext taskContext, IWorkflowContext context, CancellationToken cancellationToken)
{
@@ -319,18 +287,15 @@ internal class MagenticOrchestrator(AIAgent managerAgent, List<AIAgent> team, Ta
return;
}
string nextExecutorId = AIAgentHostExecutor.IdFor(nextAgent);
if (!string.IsNullOrWhiteSpace(taskContext.ProgressLedger.InstructionOrQuestion))
{
ChatMessage instruction = new(ChatRole.Assistant, taskContext.ProgressLedger.InstructionOrQuestion);
taskContext.ChatHistory.Add(instruction);
// Target the instruction at the chosen speaker only.
await context.SendMessageAsync(instruction, nextExecutorId, cancellationToken).ConfigureAwait(false);
await context.SendMessageAsync(instruction, cancellationToken).ConfigureAwait(false);
}
this._currentSpeakerExecutorId = nextExecutorId;
string nextExecutorId = AIAgentHostExecutor.IdFor(nextAgent);
await context.SendMessageAsync(new TurnToken(taskContext.EmitUpdateEvents), nextExecutorId, cancellationToken).ConfigureAwait(false);
}
@@ -338,7 +303,6 @@ internal class MagenticOrchestrator(AIAgent managerAgent, List<AIAgent> team, Ta
{
bool wasStalled = taskContext.IsStalled;
taskContext.Reset();
this._currentSpeakerExecutorId = null;
await context.SendMessageAsync(new ResetChatSignal(), cancellationToken: cancellationToken).ConfigureAwait(false);
await this.UpdatePlanAndDelegateAsync(taskContext, context, cancellationToken, replanAfterStall: wasStalled).ConfigureAwait(false);
@@ -349,9 +313,9 @@ internal class MagenticOrchestrator(AIAgent managerAgent, List<AIAgent> team, Ta
List<ChatMessage> messages = [await this._manager.PrepareFinalAnswerAsync(taskContext, context, cancellationToken).ConfigureAwait(false)];
await context.YieldOutputAsync(messages, cancellationToken).ConfigureAwait(false);
taskContext.IsTerminated = true;
this._currentSpeakerExecutorId = null;
}
private const string CurrentTurnEmitUpdateEventsKey = nameof(CurrentTurnEmitUpdateEventsKey);
protected internal override async ValueTask OnCheckpointingAsync(IWorkflowContext context, CancellationToken cancellationToken = default)
{
Task contextStateTask = this._taskContext == null
@@ -361,21 +325,14 @@ internal class MagenticOrchestrator(AIAgent managerAgent, List<AIAgent> team, Ta
cancellationToken: cancellationToken)
.AsTask();
Task currentSpeakerTask = context.QueueStateUpdateAsync(MagenticConstants.CurrentSpeakerStateKey,
this._currentSpeakerExecutorId,
cancellationToken: cancellationToken)
.AsTask();
await Task.WhenAll(base.OnCheckpointingAsync(context, cancellationToken).AsTask(),
contextStateTask,
currentSpeakerTask).ConfigureAwait(false);
contextStateTask).ConfigureAwait(false);
}
protected internal override async ValueTask OnCheckpointRestoredAsync(IWorkflowContext context, CancellationToken cancellationToken = default)
{
await Task.WhenAll(base.OnCheckpointRestoredAsync(context, cancellationToken).AsTask(),
LoadContextStateAsync(),
LoadCurrentSpeakerAsync()).ConfigureAwait(false);
await Task.WhenAll(base.OnCheckpointRestoredAsync(context, cancellationToken).AsTask(), LoadContextStateAsync())
.ConfigureAwait(false);
async Task LoadContextStateAsync()
{
@@ -387,11 +344,5 @@ internal class MagenticOrchestrator(AIAgent managerAgent, List<AIAgent> team, Ta
this._taskContext = new MagenticTaskContext(state, team, limits, []);
}
}
async Task LoadCurrentSpeakerAsync()
{
this._currentSpeakerExecutorId = await context.ReadStateAsync<string?>(MagenticConstants.CurrentSpeakerStateKey, cancellationToken: cancellationToken)
.ConfigureAwait(false);
}
}
}
@@ -38,8 +38,6 @@ namespace Microsoft.Agents.AI;
/// </remarks>
public sealed partial class ChatClientAgent : AIAgent
{
private const string AGUIProviderName = "ag-ui";
private readonly ChatClientAgentOptions? _agentOptions;
private readonly HashSet<string> _aiContextProviderStateKeys;
private readonly AIAgentMetadata _agentMetadata;
@@ -564,7 +562,6 @@ public sealed partial class ChatClientAgent : AIAgent
requestChatOptions.ModelId ??= this._agentOptions.ChatOptions.ModelId;
requestChatOptions.PresencePenalty ??= this._agentOptions.ChatOptions.PresencePenalty;
requestChatOptions.ResponseFormat ??= this._agentOptions.ChatOptions.ResponseFormat;
requestChatOptions.Reasoning ??= this._agentOptions.ChatOptions.Reasoning;
requestChatOptions.Seed ??= this._agentOptions.ChatOptions.Seed;
requestChatOptions.Temperature ??= this._agentOptions.ChatOptions.Temperature;
requestChatOptions.TopP ??= this._agentOptions.ChatOptions.TopP;
@@ -818,7 +815,7 @@ public sealed partial class ChatClientAgent : AIAgent
if (!string.IsNullOrWhiteSpace(responseConversationId))
{
if (!IsAGUIProviderName(this._agentMetadata.ProviderName) && this._agentOptions?.ChatHistoryProvider is not null)
if (this._agentOptions?.ChatHistoryProvider is not null)
{
// The agent has a ChatHistoryProvider configured, but the service returned a conversation id,
// meaning the service manages chat history server-side. Both cannot be used simultaneously.
@@ -932,9 +929,6 @@ public sealed partial class ChatClientAgent : AIAgent
}
}
private static bool IsAGUIProviderName(string? providerName) =>
string.Equals(providerName, AGUIProviderName, StringComparison.Ordinal);
/// <summary>
/// Ensures that <see cref="AIAgent.CurrentRunContext"/> contains the resolved session.
/// </summary>
@@ -982,17 +976,12 @@ public sealed partial class ChatClientAgent : AIAgent
private ChatHistoryProvider? ResolveChatHistoryProvider(ChatOptions? chatOptions)
{
ChatHistoryProvider? provider =
chatOptions?.ConversationId is null || IsAGUIProviderName(this._agentMetadata.ProviderName)
? this.ChatHistoryProvider
: null;
ChatHistoryProvider? provider = chatOptions?.ConversationId is null ? this.ChatHistoryProvider : null;
// If someone provided an override ChatHistoryProvider via AdditionalProperties, we should use that instead.
if (chatOptions?.AdditionalProperties?.TryGetValue(out ChatHistoryProvider? overrideProvider) is true)
{
if (!IsAGUIProviderName(this._agentMetadata.ProviderName) &&
this._agentOptions?.ThrowOnChatHistoryProviderConflict is true &&
string.IsNullOrWhiteSpace(chatOptions?.ConversationId) is false)
if (this._agentOptions?.ThrowOnChatHistoryProviderConflict is true && string.IsNullOrWhiteSpace(chatOptions?.ConversationId) is false)
{
throw new InvalidOperationException(
$"Only {nameof(ChatClientAgentSession.ConversationId)} or {nameof(this.ChatHistoryProvider)} may be used, but not both. The current {nameof(ChatClientAgentSession)} has a {nameof(ChatClientAgentSession.ConversationId)} indicating server-side chat history management, but an override {nameof(this.ChatHistoryProvider)} was provided via {nameof(AgentRunOptions.AdditionalProperties)}.");
@@ -243,46 +243,6 @@ public sealed class AGUIAgentTests
Assert.Contains(updates, u => u.Text == "Hello");
}
[Fact]
public async Task RunStreamingAsync_WithSession_SendsFullHistoryAfterThreadIdIsSetAsync()
{
// Arrange
var captureHandler = new StateCapturingTestDelegatingHandler();
captureHandler.AddResponse(
[
new RunStartedEvent { ThreadId = "thread1", RunId = "run1" },
new TextMessageStartEvent { MessageId = "msg1", Role = AGUIRoles.Assistant },
new TextMessageContentEvent { MessageId = "msg1", Delta = "First response" },
new TextMessageEndEvent { MessageId = "msg1" },
new RunFinishedEvent { ThreadId = "thread1", RunId = "run1" }
]);
captureHandler.AddResponse(
[
new RunStartedEvent { ThreadId = "thread1", RunId = "run2" },
new TextMessageStartEvent { MessageId = "msg2", Role = AGUIRoles.Assistant },
new TextMessageContentEvent { MessageId = "msg2", Delta = "Second response" },
new TextMessageEndEvent { MessageId = "msg2" },
new RunFinishedEvent { ThreadId = "thread1", RunId = "run2" }
]);
using HttpClient httpClient = new(captureHandler);
var chatClient = new AGUIChatClient(httpClient, "http://localhost/agent", null, AGUIJsonSerializerContext.Default.Options);
AIAgent agent = chatClient.AsAIAgent(instructions: null, name: "agent1", description: "Test agent", tools: []);
AgentSession session = await agent.CreateSessionAsync();
// Act
await foreach (var _ in agent.RunStreamingAsync([new ChatMessage(ChatRole.User, "First")], session))
{
}
await foreach (var _ in agent.RunStreamingAsync([new ChatMessage(ChatRole.User, "Second")], session))
{
}
// Assert
Assert.Equal([1, 3], captureHandler.CapturedMessageCounts);
}
[Fact]
public async Task DeserializeSession_WithValidState_ReturnsChatClientAgentSessionAsync()
{
@@ -1726,12 +1686,10 @@ internal sealed class CapturingTestDelegatingHandler : DelegatingHandler
internal sealed class StateCapturingTestDelegatingHandler : DelegatingHandler
{
private readonly Queue<Func<HttpRequestMessage, Task<HttpResponseMessage>>> _responseFactories = new();
private readonly List<int> _capturedMessageCounts = [];
public bool RequestWasMade { get; private set; }
public JsonElement? CapturedState { get; private set; }
public int CapturedMessageCount { get; private set; }
public IReadOnlyList<int> CapturedMessageCounts => this._capturedMessageCounts;
public void AddResponse(BaseEvent[] events)
{
@@ -1756,7 +1714,6 @@ internal sealed class StateCapturingTestDelegatingHandler : DelegatingHandler
this.CapturedState = input.State;
}
this.CapturedMessageCount = input.Messages.Count();
this._capturedMessageCounts.Add(this.CapturedMessageCount);
}
if (this._responseFactories.Count == 0)
@@ -182,7 +182,7 @@ public sealed class WorkflowConsoleAppSamplesValidation(ITestOutputHelper output
}
}
[RetryFact(2, 5000, Skip = "KeyNotFoundException in workflow execution. See https://github.com/microsoft/agent-framework/issues/6404")]
[RetryFact(2, 5000)]
public async Task WorkflowEventsSampleValidationAsync()
{
using CancellationTokenSource testTimeoutCts = this.CreateTestTimeoutCts(s_testTimeout);
@@ -278,7 +278,7 @@ public sealed class WorkflowConsoleAppSamplesValidation(ITestOutputHelper output
});
}
[RetryFact(2, 5000, Skip = "KeyNotFoundException in workflow execution. See https://github.com/microsoft/agent-framework/issues/6404")]
[RetryFact(2, 5000)]
public async Task WorkflowSharedStateSampleValidationAsync()
{
using CancellationTokenSource testTimeoutCts = this.CreateTestTimeoutCts(s_testTimeout);
@@ -376,7 +376,7 @@ public sealed class WorkflowConsoleAppSamplesValidation(ITestOutputHelper output
});
}
[RetryFact(2, 5000, Skip = "KeyNotFoundException in workflow execution. See https://github.com/microsoft/agent-framework/issues/6404")]
[RetryFact(2, 5000)]
public async Task SubWorkflowsSampleValidationAsync()
{
using CancellationTokenSource testTimeoutCts = this.CreateTestTimeoutCts(s_testTimeout);
@@ -452,7 +452,7 @@ public sealed class WorkflowConsoleAppSamplesValidation(ITestOutputHelper output
});
}
[RetryFact(2, 5000, Skip = "KeyNotFoundException in workflow execution. See https://github.com/microsoft/agent-framework/issues/6404")]
[RetryFact(2, 5000)]
public async Task WorkflowHITLSampleValidationAsync()
{
using CancellationTokenSource testTimeoutCts = this.CreateTestTimeoutCts(s_testTimeout);
@@ -43,7 +43,7 @@ public class HostedFoundryMemoryProviderScopesTests
}
[Fact]
public void PerUserAndChat_ComposesUserAndChatWithEscapedSeparator()
public void PerUserAndChat_ComposesUserAndChatWithColon()
{
// Arrange
var session = CreateTaggedSession(TestUserId, TestChatId);
@@ -54,51 +54,7 @@ public class HostedFoundryMemoryProviderScopesTests
// Assert
Assert.NotNull(state);
Assert.Equal($"{TestUserId}::{TestChatId}", state.Scope.Scope);
}
[Fact]
public void PerUserAndChat_EscapesColonsInUserAndChat()
{
// Arrange
var session = CreateTaggedSession("alice:finance", "q2:final");
var initializer = HostedFoundryMemoryProviderScopes.PerUserAndChat();
// Act
var state = initializer(session);
// Assert - colons inside each part are escaped as \: , parts joined with ::
Assert.Equal(@"alice\:finance::q2\:final", state.Scope.Scope);
}
[Fact]
public void PerUserAndChat_EscapesBackslashesInUserAndChat()
{
// Arrange
var session = CreateTaggedSession(@"alice\corp", @"chat\1");
var initializer = HostedFoundryMemoryProviderScopes.PerUserAndChat();
// Act
var state = initializer(session);
// Assert - backslashes escaped first as \\ , parts joined with ::
Assert.Equal(@"alice\\corp::chat\\1", state.Scope.Scope);
}
[Fact]
public void PerUserAndChat_DistinctContextsDoNotCollide()
{
// Arrange - two distinct (UserId, ChatId) pairs that collide under raw-colon composition.
var sessionA = CreateTaggedSession("alice:finance", "q2");
var sessionB = CreateTaggedSession("alice", "finance:q2");
var initializer = HostedFoundryMemoryProviderScopes.PerUserAndChat();
// Act
var scopeA = initializer(sessionA).Scope.Scope;
var scopeB = initializer(sessionB).Scope.Scope;
// Assert
Assert.NotEqual(scopeA, scopeB);
Assert.Equal($"{TestUserId}:{TestChatId}", state.Scope.Scope);
}
[Fact]
@@ -4,8 +4,7 @@ using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using GitHub.Copilot;
using GitHub.Copilot.Rpc;
using GitHub.Copilot.SDK;
using Microsoft.Extensions.AI;
namespace Microsoft.Agents.AI.GitHub.Copilot.IntegrationTests;
@@ -14,8 +13,8 @@ public class GitHubCopilotAgentTests
{
private const string SkipReason = "Integration tests require GitHub Copilot CLI installed. For local execution only.";
private static Task<PermissionDecision> OnPermissionRequestAsync(PermissionRequest request, PermissionInvocation invocation)
=> Task.FromResult(PermissionDecision.ApproveOnce());
private static Task<PermissionRequestResult> OnPermissionRequestAsync(PermissionRequest request, PermissionInvocation invocation)
=> Task.FromResult(new PermissionRequestResult { Kind = PermissionRequestResultKind.Approved });
[Fact(Skip = SkipReason)]
public async Task RunAsync_WithSimplePrompt_ReturnsResponseAsync()
@@ -3,7 +3,6 @@
<PropertyGroup>
<!-- GitHub.Copilot.SDK only supports .NET 8.0+ -->
<TargetFrameworks>$(TargetFrameworksCore)</TargetFrameworks>
<NoWarn>$(NoWarn);GHCP001</NoWarn>
</PropertyGroup>
<ItemGroup>
@@ -2,7 +2,7 @@
using System;
using System.Collections.Generic;
using GitHub.Copilot;
using GitHub.Copilot.SDK;
using Microsoft.Extensions.AI;
namespace Microsoft.Agents.AI.GitHub.Copilot.UnitTests;
@@ -16,7 +16,7 @@ public sealed class CopilotClientExtensionsTests
public void AsAIAgent_WithAllParameters_ReturnsGitHubCopilotAgentWithSpecifiedProperties()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
const string TestId = "test-agent-id";
const string TestName = "Test Agent";
@@ -37,7 +37,7 @@ public sealed class CopilotClientExtensionsTests
public void AsAIAgent_WithMinimalParameters_ReturnsGitHubCopilotAgent()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
// Act
var agent = copilotClient.AsAIAgent(ownsClient: false, tools: null);
@@ -61,7 +61,7 @@ public sealed class CopilotClientExtensionsTests
public void AsAIAgent_WithOwnsClient_ReturnsAgentThatOwnsClient()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
// Act
var agent = copilotClient.AsAIAgent(ownsClient: true, tools: null);
@@ -75,7 +75,7 @@ public sealed class CopilotClientExtensionsTests
public void AsAIAgent_WithTools_ReturnsAgentWithTools()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
List<AITool> tools = [AIFunctionFactory.Create(() => "test", "TestFunc", "Test function")];
// Act
@@ -3,8 +3,7 @@
using System;
using System.Collections.Generic;
using System.Threading.Tasks;
using GitHub.Copilot;
using GitHub.Copilot.Rpc;
using GitHub.Copilot.SDK;
using Microsoft.Extensions.AI;
namespace Microsoft.Agents.AI.GitHub.Copilot.UnitTests;
@@ -18,7 +17,7 @@ public sealed class GitHubCopilotAgentTests
public void Constructor_WithCopilotClient_InitializesPropertiesCorrectly()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
const string TestId = "test-id";
const string TestName = "test-name";
const string TestDescription = "test-description";
@@ -43,7 +42,7 @@ public sealed class GitHubCopilotAgentTests
public void Constructor_WithDefaultParameters_UsesBaseProperties()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
// Act
var agent = new GitHubCopilotAgent(copilotClient, ownsClient: false, tools: null);
@@ -59,7 +58,7 @@ public sealed class GitHubCopilotAgentTests
public async Task CreateSessionAsync_ReturnsGitHubCopilotAgentSessionAsync()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
var agent = new GitHubCopilotAgent(copilotClient, ownsClient: false, tools: null);
// Act
@@ -74,7 +73,7 @@ public sealed class GitHubCopilotAgentTests
public async Task CreateSessionAsync_WithSessionId_ReturnsSessionWithSessionIdAsync()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
var agent = new GitHubCopilotAgent(copilotClient, ownsClient: false, tools: null);
const string TestSessionId = "test-session-id";
@@ -91,7 +90,7 @@ public sealed class GitHubCopilotAgentTests
public void Constructor_WithTools_InitializesCorrectly()
{
// Arrange
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
List<AITool> tools = [AIFunctionFactory.Create(() => "test", "TestFunc", "Test function")];
// Act
@@ -106,12 +105,12 @@ public sealed class GitHubCopilotAgentTests
public void CopySessionConfig_CopiesAllProperties()
{
// Arrange
List<AIFunctionDeclaration> tools = [AIFunctionFactory.Create(() => "test", "TestFunc", "Test function")];
List<AIFunction> tools = [AIFunctionFactory.Create(() => "test", "TestFunc", "Test function")];
var hooks = new SessionHooks();
var infiniteSessions = new InfiniteSessionConfig();
var systemMessage = new SystemMessageConfig { Mode = SystemMessageMode.Append, Content = "Be helpful" };
Func<PermissionRequest, PermissionInvocation, Task<PermissionDecision>> permissionHandler = (_, _) => Task.FromResult(PermissionDecision.ApproveOnce());
Func<UserInputRequest, UserInputInvocation, Task<UserInputResponse>> userInputHandler = (_, _) => Task.FromResult(new UserInputResponse { Answer = "input" });
PermissionRequestHandler permissionHandler = (_, _) => Task.FromResult(new PermissionRequestResult());
UserInputHandler userInputHandler = (_, _) => Task.FromResult(new UserInputResponse { Answer = "input" });
var mcpServers = new Dictionary<string, McpServerConfig> { ["server1"] = new McpStdioServerConfig() };
var source = new SessionConfig
@@ -123,7 +122,7 @@ public sealed class GitHubCopilotAgentTests
AvailableTools = ["tool1", "tool2"],
ExcludedTools = ["tool3"],
WorkingDirectory = "/workspace",
ConfigDirectory = "/config",
ConfigDir = "/config",
Hooks = hooks,
InfiniteSessions = infiniteSessions,
OnPermissionRequest = permissionHandler,
@@ -138,15 +137,17 @@ public sealed class GitHubCopilotAgentTests
// Assert
Assert.Equal("gpt-4o", result.Model);
Assert.Equal("high", result.ReasoningEffort);
Assert.Equal(systemMessage, result.SystemMessage);
Assert.Same(tools, result.Tools);
Assert.Same(systemMessage, result.SystemMessage);
Assert.Equal(new List<string> { "tool1", "tool2" }, result.AvailableTools);
Assert.Equal(new List<string> { "tool3" }, result.ExcludedTools);
Assert.Equal("/workspace", result.WorkingDirectory);
Assert.Equal("/config", result.ConfigDirectory);
Assert.Equal("/config", result.ConfigDir);
Assert.Same(hooks, result.Hooks);
Assert.Same(infiniteSessions, result.InfiniteSessions);
Assert.Same(permissionHandler, result.OnPermissionRequest);
Assert.Same(userInputHandler, result.OnUserInputRequest);
Assert.Same(mcpServers, result.McpServers);
Assert.Equal(new List<string> { "skill1" }, result.DisabledSkills);
Assert.True(result.Streaming);
}
@@ -155,12 +156,12 @@ public sealed class GitHubCopilotAgentTests
public void CopyResumeSessionConfig_CopiesAllProperties()
{
// Arrange
List<AIFunctionDeclaration> tools = [AIFunctionFactory.Create(() => "test", "TestFunc", "Test function")];
List<AIFunction> tools = [AIFunctionFactory.Create(() => "test", "TestFunc", "Test function")];
var hooks = new SessionHooks();
var infiniteSessions = new InfiniteSessionConfig();
var systemMessage = new SystemMessageConfig { Mode = SystemMessageMode.Append, Content = "Be helpful" };
Func<PermissionRequest, PermissionInvocation, Task<PermissionDecision>> permissionHandler = (_, _) => Task.FromResult(PermissionDecision.ApproveOnce());
Func<UserInputRequest, UserInputInvocation, Task<UserInputResponse>> userInputHandler = (_, _) => Task.FromResult(new UserInputResponse { Answer = "input" });
PermissionRequestHandler permissionHandler = (_, _) => Task.FromResult(new PermissionRequestResult());
UserInputHandler userInputHandler = (_, _) => Task.FromResult(new UserInputResponse { Answer = "input" });
var mcpServers = new Dictionary<string, McpServerConfig> { ["server1"] = new McpStdioServerConfig() };
var source = new SessionConfig
@@ -172,7 +173,7 @@ public sealed class GitHubCopilotAgentTests
AvailableTools = ["tool1", "tool2"],
ExcludedTools = ["tool3"],
WorkingDirectory = "/workspace",
ConfigDirectory = "/config",
ConfigDir = "/config",
Hooks = hooks,
InfiniteSessions = infiniteSessions,
OnPermissionRequest = permissionHandler,
@@ -192,7 +193,7 @@ public sealed class GitHubCopilotAgentTests
Assert.Equal(new List<string> { "tool1", "tool2" }, result.AvailableTools);
Assert.Equal(new List<string> { "tool3" }, result.ExcludedTools);
Assert.Equal("/workspace", result.WorkingDirectory);
Assert.Equal("/config", result.ConfigDirectory);
Assert.Equal("/config", result.ConfigDir);
Assert.Same(hooks, result.Hooks);
Assert.Same(infiniteSessions, result.InfiniteSessions);
Assert.Same(permissionHandler, result.OnPermissionRequest);
@@ -217,7 +218,7 @@ public sealed class GitHubCopilotAgentTests
Assert.Null(result.OnUserInputRequest);
Assert.Null(result.Hooks);
Assert.Null(result.WorkingDirectory);
Assert.Null(result.ConfigDirectory);
Assert.Null(result.ConfigDir);
Assert.True(result.Streaming);
}
@@ -232,7 +233,7 @@ public sealed class GitHubCopilotAgentTests
Content = "Some streamed content that was already delivered via delta events"
}
};
CopilotClient copilotClient = new(new CopilotClientOptions());
CopilotClient copilotClient = new(new CopilotClientOptions { AutoStart = false });
const string TestId = "agent-id";
var agent = new GitHubCopilotAgent(copilotClient, ownsClient: false, id: TestId, tools: null);
AgentResponseUpdate result = agent.ConvertToAgentResponseUpdate(assistantMessage);
@@ -3,7 +3,6 @@
<PropertyGroup>
<!-- GitHub.Copilot.SDK only supports .NET 8.0+ -->
<TargetFrameworks>$(TargetFrameworksCore)</TargetFrameworks>
<NoWarn>$(NoWarn);GHCP001</NoWarn>
</PropertyGroup>
<ItemGroup>
@@ -27,7 +27,6 @@ public class HarnessAgentOptionsTests
Assert.Null(options.ChatHistoryProvider);
Assert.Null(options.AIContextProviders);
Assert.False(options.DisableToolApproval);
Assert.False(options.DisableNonApprovalRequiredFunctionBypassing);
Assert.False(options.DisableFileMemory);
Assert.False(options.DisableFileAccess);
Assert.False(options.DisableWebSearch);
@@ -81,7 +80,6 @@ public class HarnessAgentOptionsTests
AIContextProviders = contextProviders,
MaximumIterationsPerRequest = 42,
DisableToolApproval = true,
DisableNonApprovalRequiredFunctionBypassing = true,
DisableFileMemory = true,
FileMemoryStore = fileMemoryStore,
DisableFileAccess = true,
@@ -114,7 +112,6 @@ public class HarnessAgentOptionsTests
Assert.Same(contextProviders, options.AIContextProviders);
Assert.Equal(42, options.MaximumIterationsPerRequest);
Assert.True(options.DisableToolApproval);
Assert.True(options.DisableNonApprovalRequiredFunctionBypassing);
Assert.True(options.DisableFileMemory);
Assert.Same(fileMemoryStore, options.FileMemoryStore);
Assert.True(options.DisableFileAccess);
@@ -21,12 +21,9 @@ public class HarnessAgentTests
/// <summary>
/// Creates a HarnessAgent with all default features disabled to isolate tests for specific behaviors.
/// Compaction is enabled by default for backward compatibility with existing tests.
/// </summary>
private static HarnessAgentOptions CreateAllDisabledOptions() => new()
{
MaxContextWindowTokens = TestMaxContextWindowTokens,
MaxOutputTokens = TestMaxOutputTokens,
DisableToolApproval = true,
DisableOpenTelemetry = true,
DisableFileMemory = true,
@@ -46,7 +43,7 @@ public class HarnessAgentTests
public void Constructor_ThrowsWhenChatClientIsNull()
{
// Act & Assert
Assert.Throws<ArgumentNullException>(() => new HarnessAgent(null!));
Assert.Throws<ArgumentNullException>(() => new HarnessAgent(null!, TestMaxContextWindowTokens, TestMaxOutputTokens));
}
/// <summary>
@@ -57,10 +54,9 @@ public class HarnessAgentTests
{
// Arrange
var chatClient = new Mock<IChatClient>().Object;
var options = new HarnessAgentOptions { MaxContextWindowTokens = 0, MaxOutputTokens = TestMaxOutputTokens };
// Act & Assert
Assert.Throws<ArgumentOutOfRangeException>(() => new HarnessAgent(chatClient, options));
Assert.Throws<ArgumentOutOfRangeException>(() => new HarnessAgent(chatClient, 0, TestMaxOutputTokens));
}
/// <summary>
@@ -71,10 +67,9 @@ public class HarnessAgentTests
{
// Arrange
var chatClient = new Mock<IChatClient>().Object;
var options = new HarnessAgentOptions { MaxContextWindowTokens = 100_000, MaxOutputTokens = 100_000 };
// Act & Assert
Assert.Throws<ArgumentOutOfRangeException>(() => new HarnessAgent(chatClient, options));
Assert.Throws<ArgumentOutOfRangeException>(() => new HarnessAgent(chatClient, 100_000, 100_000));
}
/// <summary>
@@ -87,7 +82,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens);
// Assert
Assert.NotNull(agent);
@@ -110,7 +105,7 @@ public class HarnessAgentTests
options.Description = "A test agent";
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
// Assert
Assert.Equal("TestAgent", agent.Name);
@@ -129,7 +124,7 @@ public class HarnessAgentTests
options.Id = "my-agent-id";
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
// Assert
Assert.Equal("my-agent-id", agent.Id);
@@ -149,7 +144,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -169,7 +164,7 @@ public class HarnessAgentTests
options.ChatOptions = new ChatOptions { Temperature = 0.5f };
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -189,7 +184,7 @@ public class HarnessAgentTests
options.ChatOptions = new ChatOptions { Instructions = "You are a custom assistant." };
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -210,7 +205,7 @@ public class HarnessAgentTests
options.HarnessInstructions = "Custom harness rules.";
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -231,7 +226,7 @@ public class HarnessAgentTests
options.ChatOptions = new ChatOptions { Instructions = "You are a research agent." };
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -252,7 +247,7 @@ public class HarnessAgentTests
options.ChatOptions = new ChatOptions { Instructions = "Agent only instructions." };
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -272,7 +267,7 @@ public class HarnessAgentTests
options.HarnessInstructions = string.Empty;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -294,7 +289,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -315,7 +310,7 @@ public class HarnessAgentTests
options.ChatHistoryProvider = customProvider;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -337,7 +332,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -358,7 +353,7 @@ public class HarnessAgentTests
var rawClient = mockClient.Object;
// Act
var agent = new HarnessAgent(rawClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(rawClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert — the pipeline wraps the raw client, so the outer client is not the same object.
@@ -383,7 +378,7 @@ public class HarnessAgentTests
options.AIContextProviders = [customProvider];
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert — the custom provider should appear in the inner agent's AIContextProviders.
@@ -403,7 +398,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -437,7 +432,7 @@ public class HarnessAgentTests
var options = CreateAllDisabledOptions();
options.ChatOptions = new ChatOptions { Tools = [tool] };
var agent = new HarnessAgent(mockClient.Object, options);
var agent = new HarnessAgent(mockClient.Object, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var session = await agent.CreateSessionAsync();
// Act
@@ -464,10 +459,8 @@ public class HarnessAgentTests
};
// Act
_ = new HarnessAgent(chatClient, new HarnessAgentOptions
_ = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, new HarnessAgentOptions
{
MaxContextWindowTokens = TestMaxContextWindowTokens,
MaxOutputTokens = TestMaxOutputTokens,
ChatOptions = sourceChatOptions,
});
@@ -490,7 +483,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
// Assert
Assert.Same(agent, agent.GetService<HarnessAgent>());
@@ -506,7 +499,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
// Assert
Assert.NotNull(agent.GetService<ChatClientAgent>());
@@ -531,7 +524,7 @@ public class HarnessAgentTests
It.IsAny<CancellationToken>()))
.ReturnsAsync(new ChatResponse(new ChatMessage(ChatRole.Assistant, "Hello!")));
var agent = new HarnessAgent(mockClient.Object, CreateAllDisabledOptions());
var agent = new HarnessAgent(mockClient.Object, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var session = await agent.CreateSessionAsync();
// Act
@@ -572,7 +565,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = chatClient.AsHarnessAgent();
var agent = chatClient.AsHarnessAgent(TestMaxContextWindowTokens, TestMaxOutputTokens);
// Assert
Assert.NotNull(agent);
@@ -593,7 +586,7 @@ public class HarnessAgentTests
options.ChatOptions = new ChatOptions { Instructions = "Custom instructions" };
// Act
var agent = chatClient.AsHarnessAgent(options);
var agent = chatClient.AsHarnessAgent(TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -610,7 +603,7 @@ public class HarnessAgentTests
public void AsHarnessAgent_ThrowsWhenChatClientIsNull()
{
// Act & Assert
Assert.Throws<ArgumentNullException>(() => ((IChatClient)null!).AsHarnessAgent());
Assert.Throws<ArgumentNullException>(() => ((IChatClient)null!).AsHarnessAgent(TestMaxContextWindowTokens, TestMaxOutputTokens));
}
#endregion
@@ -629,7 +622,7 @@ public class HarnessAgentTests
options.DisableToolApproval = false;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
// Assert
Assert.NotNull(agent.GetService<ToolApprovalAgent>());
@@ -645,7 +638,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
// Assert
Assert.Null(agent.GetService<ToolApprovalAgent>());
@@ -685,7 +678,7 @@ public class HarnessAgentTests
AutoApprovalRules = [fcc => new ValueTask<bool>(fcc.Name == "ReadTool")]
};
var agent = new HarnessAgent(mockClient.Object, options);
var agent = new HarnessAgent(mockClient.Object, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var session = await agent.CreateSessionAsync();
// Act
@@ -698,97 +691,6 @@ public class HarnessAgentTests
#endregion
#region Feature: NonApprovalRequiredFunctionBypassing
/// <summary>
/// Verify that by default, when a response contains a mix of tools that require approval and tools that do not,
/// only the approval-required tool is surfaced to the caller. The non-approval-required tool is bypassed
/// (stored as auto-approved) by the <c>NonApprovalRequiredFunctionBypassingChatClient</c> decorator.
/// </summary>
[Fact]
public async Task NonApprovalRequiredFunctionBypassing_BypassesNonApprovalToolsByDefaultAsync()
{
// Arrange — the model requests both a normal tool and an approval-required tool in the same turn.
var normalTool = AIFunctionFactory.Create(() => "result", "NormalTool");
var approvalTool = new ApprovalRequiredAIFunction(AIFunctionFactory.Create(() => "result", "ApprovalTool"));
var mockClient = new Mock<IChatClient>();
mockClient
.Setup(c => c.GetResponseAsync(
It.IsAny<IEnumerable<ChatMessage>>(),
It.IsAny<ChatOptions>(),
It.IsAny<CancellationToken>()))
.ReturnsAsync(() => new ChatResponse(new ChatMessage(ChatRole.Assistant,
[
new FunctionCallContent("call1", "NormalTool"),
new FunctionCallContent("call2", "ApprovalTool"),
])));
// Disable ToolApproval so the approval requests surface in the response instead of being handled.
var options = CreateAllDisabledOptions();
options.ChatOptions = new ChatOptions { Tools = [normalTool, approvalTool] };
var agent = new HarnessAgent(mockClient.Object, options);
var session = await agent.CreateSessionAsync();
// Act
var response = await agent.RunAsync([new ChatMessage(ChatRole.User, "Hi")], session);
// Assert — only the approval-required tool surfaces as an approval request; the normal tool is bypassed.
var approvalRequests = response.Messages
.SelectMany(m => m.Contents)
.OfType<ToolApprovalRequestContent>()
.ToList();
var approvalRequest = Assert.Single(approvalRequests);
Assert.Equal("ApprovalTool", Assert.IsType<FunctionCallContent>(approvalRequest.ToolCall).Name);
}
/// <summary>
/// Verify that when bypassing is disabled, all tools (including those that do not require approval) are surfaced
/// as approval requests, reflecting the all-or-nothing behavior of <see cref="FunctionInvokingChatClient"/>.
/// </summary>
[Fact]
public async Task NonApprovalRequiredFunctionBypassing_SurfacesAllApprovalsWhenDisabledAsync()
{
// Arrange — the model requests both a normal tool and an approval-required tool in the same turn.
var normalTool = AIFunctionFactory.Create(() => "result", "NormalTool");
var approvalTool = new ApprovalRequiredAIFunction(AIFunctionFactory.Create(() => "result", "ApprovalTool"));
var mockClient = new Mock<IChatClient>();
mockClient
.Setup(c => c.GetResponseAsync(
It.IsAny<IEnumerable<ChatMessage>>(),
It.IsAny<ChatOptions>(),
It.IsAny<CancellationToken>()))
.ReturnsAsync(() => new ChatResponse(new ChatMessage(ChatRole.Assistant,
[
new FunctionCallContent("call1", "NormalTool"),
new FunctionCallContent("call2", "ApprovalTool"),
])));
var options = CreateAllDisabledOptions();
options.DisableNonApprovalRequiredFunctionBypassing = true;
options.ChatOptions = new ChatOptions { Tools = [normalTool, approvalTool] };
var agent = new HarnessAgent(mockClient.Object, options);
var session = await agent.CreateSessionAsync();
// Act
var response = await agent.RunAsync([new ChatMessage(ChatRole.User, "Hi")], session);
// Assert — both tools surface as approval requests because bypassing is disabled.
var approvalRequests = response.Messages
.SelectMany(m => m.Contents)
.OfType<ToolApprovalRequestContent>()
.Select(r => ((FunctionCallContent)r.ToolCall).Name)
.ToList();
Assert.Equal(2, approvalRequests.Count);
Assert.Contains("NormalTool", approvalRequests);
Assert.Contains("ApprovalTool", approvalRequests);
}
#endregion
#region Feature: OpenTelemetry
/// <summary>
@@ -803,7 +705,7 @@ public class HarnessAgentTests
options.DisableOpenTelemetry = false;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
// Assert
Assert.NotNull(agent.GetService<OpenTelemetryAgent>());
@@ -819,7 +721,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
// Assert
Assert.Null(agent.GetService<OpenTelemetryAgent>());
@@ -838,7 +740,7 @@ public class HarnessAgentTests
options.OpenTelemetrySourceName = "MyApp.AgentTracing";
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
// Assert
Assert.NotNull(agent.GetService<OpenTelemetryAgent>());
@@ -865,7 +767,7 @@ public class HarnessAgentTests
var options = CreateAllDisabledOptions();
options.DisableWebSearch = false;
var agent = new HarnessAgent(mockClient.Object, options);
var agent = new HarnessAgent(mockClient.Object, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var session = await agent.CreateSessionAsync();
// Act
@@ -890,7 +792,7 @@ public class HarnessAgentTests
.Callback<IEnumerable<ChatMessage>, ChatOptions?, CancellationToken>((_, opts, _) => capturedOptions = opts)
.ReturnsAsync(new ChatResponse(new ChatMessage(ChatRole.Assistant, "Done")));
var agent = new HarnessAgent(mockClient.Object, CreateAllDisabledOptions());
var agent = new HarnessAgent(mockClient.Object, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var session = await agent.CreateSessionAsync();
// Act
@@ -923,7 +825,7 @@ public class HarnessAgentTests
options.DisableWebSearch = false;
options.ChatOptions = new ChatOptions { Tools = [userTool] };
var agent = new HarnessAgent(mockClient.Object, options);
var agent = new HarnessAgent(mockClient.Object, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var session = await agent.CreateSessionAsync();
// Act
@@ -951,7 +853,7 @@ public class HarnessAgentTests
options.DisableTodoProvider = false;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -969,7 +871,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -996,7 +898,7 @@ public class HarnessAgentTests
options.DisableAgentModeProvider = false;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1014,7 +916,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1045,7 +947,7 @@ public class HarnessAgentTests
};
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert — AgentModeProvider should be present (we can't easily inspect its internal options,
@@ -1070,7 +972,7 @@ public class HarnessAgentTests
options.DisableFileMemory = false;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1088,7 +990,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1113,7 +1015,7 @@ public class HarnessAgentTests
options.FileMemoryStore = customStore;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert — FileMemoryProvider should be present with the custom store.
@@ -1137,7 +1039,7 @@ public class HarnessAgentTests
options.DisableFileAccess = false;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1155,7 +1057,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1180,7 +1082,7 @@ public class HarnessAgentTests
options.FileAccessStore = customStore;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert — FileAccessProvider should be present with the custom store.
@@ -1204,7 +1106,7 @@ public class HarnessAgentTests
options.DisableAgentSkillsProvider = false;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1222,7 +1124,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1247,7 +1149,7 @@ public class HarnessAgentTests
options.AgentSkillsSource = customSource;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert — AgentSkillsProvider should be present.
@@ -1271,7 +1173,7 @@ public class HarnessAgentTests
options.MaximumIterationsPerRequest = 42;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
var ficc = innerAgent!.ChatClient.GetService<FunctionInvokingChatClient>();
@@ -1290,7 +1192,7 @@ public class HarnessAgentTests
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions());
var innerAgent = agent.GetService<ChatClientAgent>();
var ficc = innerAgent!.ChatClient.GetService<FunctionInvokingChatClient>();
@@ -1318,7 +1220,7 @@ public class HarnessAgentTests
.ReturnsAsync(new ChatResponse(new ChatMessage(ChatRole.Assistant, "Done")));
// Act
var agent = new HarnessAgent(mockClient.Object);
var agent = new HarnessAgent(mockClient.Object, TestMaxContextWindowTokens, TestMaxOutputTokens);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert — agent wrappers
@@ -1361,7 +1263,7 @@ public class HarnessAgentTests
options.BackgroundAgents = [bgAgentMock.Object];
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1381,7 +1283,7 @@ public class HarnessAgentTests
options.BackgroundAgents = null;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1404,7 +1306,7 @@ public class HarnessAgentTests
options.BackgroundAgents = Array.Empty<AIAgent>();
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1435,7 +1337,7 @@ public class HarnessAgentTests
options.BackgroundAgentsProviderOptions = providerOptions;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
var bgProvider = innerAgent!.AIContextProviders!.OfType<BackgroundAgentsProvider>().Single();
@@ -1472,7 +1374,7 @@ public class HarnessAgentTests
options.BackgroundAgents = [agent1Mock.Object, agent2Mock.Object];
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
var bgProvider = innerAgent!.AIContextProviders!.OfType<BackgroundAgentsProvider>().Single();
@@ -1513,7 +1415,7 @@ public class HarnessAgentTests
options.ShellExecutor = executorMock.Object;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1533,7 +1435,7 @@ public class HarnessAgentTests
options.ShellExecutor = null;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert
@@ -1565,7 +1467,7 @@ public class HarnessAgentTests
options.ShellExecutor = executorMock.Object;
// Act
var agent = new HarnessAgent(chatClientMock.Object, options);
var agent = new HarnessAgent(chatClientMock.Object, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var session = await agent.CreateSessionAsync();
await agent.RunAsync([new ChatMessage(ChatRole.User, "Hi")], session);
@@ -1594,7 +1496,7 @@ public class HarnessAgentTests
options.ShellEnvironmentProviderOptions = envOptions;
// Act
var agent = new HarnessAgent(chatClient, options);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options);
var innerAgent = agent.GetService<ChatClientAgent>();
// Assert — provider should exist (options wiring is validated by the provider's behavior)
@@ -1618,7 +1520,7 @@ public class HarnessAgentTests
var loggerFactory = new Mock<ILoggerFactory>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions(), loggerFactory);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions(), loggerFactory);
// Assert
Assert.NotNull(agent);
@@ -1635,7 +1537,7 @@ public class HarnessAgentTests
var services = new Mock<IServiceProvider>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions(), services: services);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions(), services: services);
// Assert
Assert.NotNull(agent);
@@ -1653,7 +1555,7 @@ public class HarnessAgentTests
var services = new Mock<IServiceProvider>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions(), loggerFactory, services);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions(), loggerFactory, services);
// Assert
Assert.NotNull(agent);
@@ -1671,7 +1573,7 @@ public class HarnessAgentTests
var services = new Mock<IServiceProvider>().Object;
// Act
var agent = chatClient.AsHarnessAgent(CreateAllDisabledOptions(), loggerFactory, services);
var agent = chatClient.AsHarnessAgent(TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions(), loggerFactory, services);
// Assert
Assert.NotNull(agent);
@@ -1693,8 +1595,6 @@ public class HarnessAgentTests
// Act — use options that leave CompactionProvider and AgentSkillsProvider enabled
var options = new HarnessAgentOptions
{
MaxContextWindowTokens = TestMaxContextWindowTokens,
MaxOutputTokens = TestMaxOutputTokens,
DisableToolApproval = true,
DisableOpenTelemetry = true,
DisableFileMemory = true,
@@ -1703,7 +1603,7 @@ public class HarnessAgentTests
DisableTodoProvider = true,
DisableAgentModeProvider = true,
};
var agent = new HarnessAgent(chatClient, options, mockLoggerFactory.Object);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, options, mockLoggerFactory.Object);
// Assert — CreateLogger should have been called by one or more downstream components
Assert.NotNull(agent);
@@ -1725,7 +1625,7 @@ public class HarnessAgentTests
.Returns(null!);
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions(), services: mockServices.Object);
var agent = new HarnessAgent(chatClient, TestMaxContextWindowTokens, TestMaxOutputTokens, CreateAllDisabledOptions(), services: mockServices.Object);
// Assert — the service provider should have been queried during pipeline construction
Assert.NotNull(agent);
@@ -1733,91 +1633,4 @@ public class HarnessAgentTests
}
#endregion
#region Compaction Opt-in
/// <summary>
/// Verify that constructing without token values succeeds (compaction disabled).
/// </summary>
[Fact]
public void Constructor_SucceedsWithoutTokenValues()
{
// Arrange
var chatClient = new Mock<IChatClient>().Object;
var options = new HarnessAgentOptions
{
DisableToolApproval = true,
DisableOpenTelemetry = true,
DisableFileMemory = true,
DisableFileAccess = true,
DisableWebSearch = true,
DisableTodoProvider = true,
DisableAgentModeProvider = true,
DisableAgentSkillsProvider = true,
};
// Act
var agent = new HarnessAgent(chatClient, options);
// Assert — compaction should be disabled (no chat reducer)
var innerAgent = agent.GetService<ChatClientAgent>();
Assert.NotNull(innerAgent);
var historyProvider = innerAgent!.ChatHistoryProvider as InMemoryChatHistoryProvider;
Assert.NotNull(historyProvider);
Assert.Null(historyProvider!.ChatReducer);
}
/// <summary>
/// Verify that when only MaxContextWindowTokens is provided (no MaxOutputTokens), compaction is disabled.
/// </summary>
[Fact]
public void Constructor_SucceedsWithOnlyMaxContextWindowTokens()
{
// Arrange
var chatClient = new Mock<IChatClient>().Object;
var options = new HarnessAgentOptions
{
MaxContextWindowTokens = TestMaxContextWindowTokens,
DisableToolApproval = true,
DisableOpenTelemetry = true,
DisableFileMemory = true,
DisableFileAccess = true,
DisableWebSearch = true,
DisableTodoProvider = true,
DisableAgentModeProvider = true,
DisableAgentSkillsProvider = true,
};
// Act
var agent = new HarnessAgent(chatClient, options);
// Assert — compaction should be disabled (only one token value provided)
var innerAgent = agent.GetService<ChatClientAgent>();
Assert.NotNull(innerAgent);
var historyProvider = innerAgent!.ChatHistoryProvider as InMemoryChatHistoryProvider;
Assert.NotNull(historyProvider);
Assert.Null(historyProvider!.ChatReducer);
}
/// <summary>
/// Verify that when both token values are provided, the agent is constructed successfully with compaction.
/// </summary>
[Fact]
public void Constructor_SucceedsWithBothTokenValues()
{
// Arrange
var chatClient = new Mock<IChatClient>().Object;
// Act
var agent = new HarnessAgent(chatClient, CreateAllDisabledOptions());
// Assert — compaction should be enabled (chat reducer configured)
var innerAgent = agent.GetService<ChatClientAgent>();
Assert.NotNull(innerAgent);
var historyProvider = innerAgent!.ChatHistoryProvider as InMemoryChatHistoryProvider;
Assert.NotNull(historyProvider);
Assert.NotNull(historyProvider!.ChatReducer);
}
#endregion
}
@@ -60,7 +60,7 @@ public sealed class SamplesValidation(ITestOutputHelper outputHelper) : IAsyncLi
await Task.CompletedTask;
}
[RetryFact(2, 5000, Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[RetryFact(2, 5000)]
public async Task SingleAgentSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "01_SingleAgent");
@@ -148,7 +148,7 @@ public sealed class SamplesValidation(ITestOutputHelper outputHelper) : IAsyncLi
});
}
[RetryFact(2, 5000, Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[RetryFact(2, 5000)]
public async Task MultiAgentOrchestrationConcurrentSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "03_AgentOrchestration_Concurrency");
@@ -198,7 +198,7 @@ public sealed class SamplesValidation(ITestOutputHelper outputHelper) : IAsyncLi
});
}
[RetryFact(2, 5000, Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[RetryFact(2, 5000)]
public async Task MultiAgentOrchestrationConditionalsSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "04_AgentOrchestration_Conditionals");
@@ -216,7 +216,7 @@ public sealed class SamplesValidation(ITestOutputHelper outputHelper) : IAsyncLi
});
}
[RetryFact(2, 5000, Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[RetryFact(2, 5000)]
public async Task SingleAgentOrchestrationHITLSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "05_AgentOrchestration_HITL");
@@ -272,7 +272,7 @@ public sealed class SamplesValidation(ITestOutputHelper outputHelper) : IAsyncLi
});
}
[RetryFact(2, 5000, Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[RetryFact(2, 5000)]
public async Task LongRunningToolsSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "06_LongRunningTools");
@@ -362,7 +362,7 @@ public sealed class SamplesValidation(ITestOutputHelper outputHelper) : IAsyncLi
});
}
[RetryFact(2, 5000, Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[RetryFact(2, 5000)]
public async Task AgentAsMcpToolAsync()
{
string samplePath = Path.Combine(s_samplesPath, "07_AgentAsMcpTool");
@@ -402,7 +402,7 @@ public sealed class SamplesValidation(ITestOutputHelper outputHelper) : IAsyncLi
});
}
[RetryFact(2, 5000, Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[RetryFact(2, 5000)]
public async Task ReliableStreamingSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "08_ReliableStreaming");
@@ -62,7 +62,7 @@ public sealed class WorkflowSamplesValidation(ITestOutputHelper outputHelper) :
return default;
}
[Fact(Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[Fact]
public async Task SequentialWorkflowSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "01_SequentialWorkflow");
@@ -168,7 +168,7 @@ public sealed class WorkflowSamplesValidation(ITestOutputHelper outputHelper) :
});
}
[Fact(Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[Fact]
public async Task HITLWorkflowSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "03_WorkflowHITL");
@@ -277,7 +277,7 @@ public sealed class WorkflowSamplesValidation(ITestOutputHelper outputHelper) :
});
}
[Fact(Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[Fact]
public async Task WorkflowMcpToolSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "04_WorkflowMcpTool");
@@ -333,7 +333,7 @@ public sealed class WorkflowSamplesValidation(ITestOutputHelper outputHelper) :
});
}
[Fact(Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[Fact]
public async Task WorkflowAndAgentsSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "05_WorkflowAndAgents");
@@ -385,7 +385,7 @@ public sealed class WorkflowSamplesValidation(ITestOutputHelper outputHelper) :
});
}
[Fact(Skip = "Azure Functions Core Tools v4 cannot auto-detect worker runtime in CI. See https://github.com/microsoft/agent-framework/issues/6402")]
[Fact]
public async Task ConcurrentWorkflowSampleValidationAsync()
{
string samplePath = Path.Combine(s_samplesPath, "02_ConcurrentWorkflow");
@@ -16,7 +16,6 @@ public class ClaimsIdentitySessionIsolationKeyProviderTests
private const string TestUserId = "test-user-id";
private const string CustomClaimType = "custom-claim-type";
private const string CustomClaimValue = "custom-claim-value";
private const string TestAuthenticationType = "TestAuth";
private readonly Mock<IHttpContextAccessor> _httpContextAccessorMock;
@@ -102,25 +101,6 @@ public class ClaimsIdentitySessionIsolationKeyProviderTests
public async Task GetSessionIsolationKeyAsyncExtractsDefaultClaimTypeAsync()
{
// Arrange
this.SetupHttpContextWithClaim(ClaimTypes.NameIdentifier, TestUserId);
var provider = new ClaimsIdentitySessionIsolationKeyProvider(this._httpContextAccessorMock.Object);
// Act
string? result = await provider.GetSessionIsolationKeyAsync();
// Assert
Assert.Equal(TestUserId, result);
}
/// <summary>
/// Verify that the default claim type is the stable, unique NameIdentifier claim rather than the
/// non-unique display name claim. This guards against the session-isolation collision described in
/// the security report where two principals sharing the same name claim received the same key.
/// </summary>
[Fact]
public async Task GetSessionIsolationKeyAsyncIgnoresNameClaimByDefaultAsync()
{
// Arrange - only a display-name claim is present; the default provider must not use it.
this.SetupHttpContextWithClaim(ClaimsIdentity.DefaultNameClaimType, TestUserId);
var provider = new ClaimsIdentitySessionIsolationKeyProvider(this._httpContextAccessorMock.Object);
@@ -128,7 +108,7 @@ public class ClaimsIdentitySessionIsolationKeyProviderTests
string? result = await provider.GetSessionIsolationKeyAsync();
// Assert
Assert.Null(result);
Assert.Equal(TestUserId, result);
}
/// <summary>
@@ -211,10 +191,10 @@ public class ClaimsIdentitySessionIsolationKeyProviderTests
const string SecondValue = "second-value";
var claims = new[]
{
new Claim(ClaimTypes.NameIdentifier, FirstValue),
new Claim(ClaimTypes.NameIdentifier, SecondValue),
new Claim(ClaimsIdentity.DefaultNameClaimType, FirstValue),
new Claim(ClaimsIdentity.DefaultNameClaimType, SecondValue),
};
var identity = new ClaimsIdentity(claims, TestAuthenticationType);
var identity = new ClaimsIdentity(claims);
var principal = new ClaimsPrincipal(identity);
var httpContext = new DefaultHttpContext
@@ -239,7 +219,7 @@ public class ClaimsIdentitySessionIsolationKeyProviderTests
public async Task GetSessionIsolationKeyAsyncHandlesEmptyClaimValueAsync()
{
// Arrange
this.SetupHttpContextWithClaim(ClaimTypes.NameIdentifier, string.Empty);
this.SetupHttpContextWithClaim(ClaimsIdentity.DefaultNameClaimType, string.Empty);
var provider = new ClaimsIdentitySessionIsolationKeyProvider(this._httpContextAccessorMock.Object);
// Act
@@ -249,66 +229,6 @@ public class ClaimsIdentitySessionIsolationKeyProviderTests
Assert.Equal(string.Empty, result);
}
/// <summary>
/// Regression test for the session-isolation collision security report: two distinct authenticated
/// principals that share the same display-name claim but have different stable identifiers and tenants
/// must produce distinct isolation keys under the default options.
/// </summary>
[Fact]
public async Task GetSessionIsolationKeyAsyncDistinctForPrincipalsSharingNameClaimAsync()
{
// Arrange - both principals share the same name claim but differ by NameIdentifier and tenant.
const string CommonName = "John Doe";
var principalA = CreatePrincipal(
new Claim(ClaimsIdentity.DefaultNameClaimType, CommonName),
new Claim(ClaimTypes.NameIdentifier, "oid-user-a"),
new Claim("http://schemas.microsoft.com/identity/claims/tenantid", "tenant-a"));
var principalB = CreatePrincipal(
new Claim(ClaimsIdentity.DefaultNameClaimType, CommonName),
new Claim(ClaimTypes.NameIdentifier, "oid-user-b"),
new Claim("http://schemas.microsoft.com/identity/claims/tenantid", "tenant-b"));
var provider = new ClaimsIdentitySessionIsolationKeyProvider(this._httpContextAccessorMock.Object);
// Act
this._httpContextAccessorMock.Setup(x => x.HttpContext).Returns(new DefaultHttpContext { User = principalA });
string? principalAKey = await provider.GetSessionIsolationKeyAsync();
this._httpContextAccessorMock.Setup(x => x.HttpContext).Returns(new DefaultHttpContext { User = principalB });
string? principalBKey = await provider.GetSessionIsolationKeyAsync();
// Assert
Assert.Equal("oid-user-a", principalAKey);
Assert.Equal("oid-user-b", principalBKey);
Assert.NotEqual(principalAKey, principalBKey);
}
/// <summary>
/// Verify that GetSessionIsolationKeyAsync returns null when the request's user is not authenticated,
/// even if a claim of the configured type is present. The provider must not derive an isolation key
/// from claims on an unauthenticated identity.
/// </summary>
[Fact]
public async Task GetSessionIsolationKeyAsyncReturnsNullWhenUserNotAuthenticatedAsync()
{
// Arrange - identity has the claim but no authentication type, so IsAuthenticated is false.
var claims = new[] { new Claim(ClaimTypes.NameIdentifier, TestUserId) };
var unauthenticatedIdentity = new ClaimsIdentity(claims);
var principal = new ClaimsPrincipal(unauthenticatedIdentity);
var httpContext = new DefaultHttpContext { User = principal };
this._httpContextAccessorMock.Setup(x => x.HttpContext).Returns(httpContext);
var provider = new ClaimsIdentitySessionIsolationKeyProvider(this._httpContextAccessorMock.Object);
// Act
string? result = await provider.GetSessionIsolationKeyAsync();
// Assert
Assert.False(unauthenticatedIdentity.IsAuthenticated);
Assert.Null(result);
}
#endregion
#region Helper Methods
@@ -316,7 +236,7 @@ public class ClaimsIdentitySessionIsolationKeyProviderTests
private void SetupHttpContextWithClaim(string claimType, string claimValue)
{
var claims = new[] { new Claim(claimType, claimValue) };
var identity = new ClaimsIdentity(claims, TestAuthenticationType);
var identity = new ClaimsIdentity(claims);
var principal = new ClaimsPrincipal(identity);
var httpContext = new DefaultHttpContext
@@ -327,8 +247,5 @@ public class ClaimsIdentitySessionIsolationKeyProviderTests
this._httpContextAccessorMock.Setup(x => x.HttpContext).Returns(httpContext);
}
private static ClaimsPrincipal CreatePrincipal(params Claim[] claims)
=> new(new ClaimsIdentity(claims, TestAuthenticationType));
#endregion
}
@@ -347,115 +347,6 @@ public class ChatClientAgent_ChatOptionsMergingTests
Assert.Equal(expectedSetting, capturedChatOptions.RawRepresentationFactory(null!));
}
/// <summary>
/// Verify that <see cref="ChatOptions.Reasoning"/> from the request takes priority over the agent's.
/// </summary>
[Fact]
public async Task ChatOptionsMergingUsesRequestReasoningOverAgentReasoningAsync()
{
// Arrange
var agentReasoning = new ReasoningOptions { Effort = ReasoningEffort.Low, Output = ReasoningOutput.Full };
var requestReasoning = new ReasoningOptions { Effort = ReasoningEffort.High, Output = ReasoningOutput.Full };
Mock<IChatClient> mockService = new();
ChatOptions? capturedChatOptions = null;
mockService.Setup(
s => s.GetResponseAsync(
It.IsAny<IEnumerable<ChatMessage>>(),
It.IsAny<ChatOptions>(),
It.IsAny<CancellationToken>()))
.Callback<IEnumerable<ChatMessage>, ChatOptions, CancellationToken>((msgs, opts, ct) =>
capturedChatOptions = opts)
.ReturnsAsync(new ChatResponse([new(ChatRole.Assistant, "response")]));
ChatClientAgent agent = new(mockService.Object, options: new()
{
ChatOptions = new ChatOptions { Reasoning = agentReasoning }
});
var messages = new List<ChatMessage> { new(ChatRole.User, "test") };
// Act
await agent.RunAsync(messages, options: new ChatClientAgentRunOptions(new ChatOptions { Reasoning = requestReasoning }));
// Assert
Assert.NotNull(capturedChatOptions);
Assert.NotNull(capturedChatOptions.Reasoning);
Assert.Equal(requestReasoning.Effort, capturedChatOptions.Reasoning.Effort);
Assert.Equal(requestReasoning.Output, capturedChatOptions.Reasoning.Output);
}
/// <summary>
/// Verify that <see cref="ChatOptions.Reasoning"/> falls back to the agent's when the request has none.
/// </summary>
[Fact]
public async Task ChatOptionsMergingFallsBackToAgentReasoningWhenRequestHasNoneAsync()
{
// Arrange
var agentReasoning = new ReasoningOptions { Effort = ReasoningEffort.Low, Output = ReasoningOutput.Full };
Mock<IChatClient> mockService = new();
ChatOptions? capturedChatOptions = null;
mockService.Setup(
s => s.GetResponseAsync(
It.IsAny<IEnumerable<ChatMessage>>(),
It.IsAny<ChatOptions>(),
It.IsAny<CancellationToken>()))
.Callback<IEnumerable<ChatMessage>, ChatOptions, CancellationToken>((msgs, opts, ct) =>
capturedChatOptions = opts)
.ReturnsAsync(new ChatResponse([new(ChatRole.Assistant, "response")]));
ChatClientAgent agent = new(mockService.Object, options: new()
{
ChatOptions = new ChatOptions { Reasoning = agentReasoning }
});
var messages = new List<ChatMessage> { new(ChatRole.User, "test") };
// Act
await agent.RunAsync(messages, options: new ChatClientAgentRunOptions(new ChatOptions()));
// Assert
Assert.NotNull(capturedChatOptions);
Assert.NotNull(capturedChatOptions.Reasoning);
Assert.Equal(agentReasoning.Effort, capturedChatOptions.Reasoning.Effort);
Assert.Equal(agentReasoning.Output, capturedChatOptions.Reasoning.Output);
}
/// <summary>
/// Verify that <see cref="ChatOptions.Reasoning"/> from the request is used when the agent has none.
/// </summary>
[Fact]
public async Task ChatOptionsMergingUsesRequestReasoningWhenAgentHasNoneAsync()
{
// Arrange
var requestReasoning = new ReasoningOptions { Effort = ReasoningEffort.High, Output = ReasoningOutput.Full };
Mock<IChatClient> mockService = new();
ChatOptions? capturedChatOptions = null;
mockService.Setup(
s => s.GetResponseAsync(
It.IsAny<IEnumerable<ChatMessage>>(),
It.IsAny<ChatOptions>(),
It.IsAny<CancellationToken>()))
.Callback<IEnumerable<ChatMessage>, ChatOptions, CancellationToken>((msgs, opts, ct) =>
capturedChatOptions = opts)
.ReturnsAsync(new ChatResponse([new(ChatRole.Assistant, "response")]));
ChatClientAgent agent = new(mockService.Object, options: new()
{
ChatOptions = new ChatOptions()
});
var messages = new List<ChatMessage> { new(ChatRole.User, "test") };
// Act
await agent.RunAsync(messages, options: new ChatClientAgentRunOptions(new ChatOptions { Reasoning = requestReasoning }));
// Assert
Assert.NotNull(capturedChatOptions);
Assert.NotNull(capturedChatOptions.Reasoning);
Assert.Equal(requestReasoning.Effort, capturedChatOptions.Reasoning.Effort);
Assert.Equal(requestReasoning.Output, capturedChatOptions.Reasoning.Output);
}
/// <summary>
/// Verify that ChatOptions merging handles all scalar properties correctly.
/// </summary>
@@ -170,67 +170,6 @@ public sealed class ForeachExecutorTest(ITestOutputHelper output) : WorkflowActi
Assert.Equal("Engineer", currentValue.GetField("role").ToObject());
}
/// <summary>
/// Power Fx wraps scalar array literals such as <c>=[1, 2, 3]</c> as <c>Table({Value: 1}, ...)</c>;
/// the loop value must expose the bare scalar, not the single-column wrapper record.
/// </summary>
[Fact]
public async Task ForeachTakeNextWithSingleColumnValueRecordAsync()
{
// Arrange
const string CurrentValueName = "CurrentValue";
this.SetVariableState(CurrentValueName);
TableDataValue tableValue = DataValue.TableFromRecords(
DataValue.RecordFromFields(new KeyValuePair<string, DataValue>("Value", new NumberDataValue(1))),
DataValue.RecordFromFields(new KeyValuePair<string, DataValue>("Value", new NumberDataValue(2))),
DataValue.RecordFromFields(new KeyValuePair<string, DataValue>("Value", new NumberDataValue(3))));
Foreach model = this.CreateModel(
displayName: nameof(ForeachTakeNextWithSingleColumnValueRecordAsync),
items: ValueExpression.Literal(tableValue),
valueName: CurrentValueName,
indexName: null);
ForeachExecutor action = new(model, this.State);
// Act
await this.ExecuteAsync(action, ForeachExecutor.Steps.Next(action.Id), action.TakeNextAsync);
// Assert
FormulaValue currentValue = this.State.Get(CurrentValueName);
Assert.IsNotType<RecordValue>(currentValue, exactMatch: false);
Assert.Equal(1m, currentValue.ToObject());
}
/// <summary>
/// Single-field records whose only field is NOT named <c>Value</c> are not Power Fx auto-wraps;
/// they are preserved as records so the field name remains accessible inside the loop body.
/// </summary>
[Fact]
public async Task ForeachTakeNextWithSingleFieldNonValueRecordAsync()
{
// Arrange
const string CurrentValueName = "CurrentValue";
this.SetVariableState(CurrentValueName);
TableDataValue tableValue = DataValue.TableFromRecords(
DataValue.RecordFromFields(new KeyValuePair<string, DataValue>("name", new StringDataValue("Alice"))));
Foreach model = this.CreateModel(
displayName: nameof(ForeachTakeNextWithSingleFieldNonValueRecordAsync),
items: ValueExpression.Literal(tableValue),
valueName: CurrentValueName,
indexName: null);
ForeachExecutor action = new(model, this.State);
// Act
await this.ExecuteAsync(action, ForeachExecutor.Steps.Next(action.Id), action.TakeNextAsync);
// Assert
RecordValue currentValue = Assert.IsType<RecordValue>(this.State.Get(CurrentValueName), exactMatch: false);
Assert.Equal("Alice", currentValue.GetField("name").ToObject());
}
[Fact]
public async Task ForeachTakeLastAsync()
{
@@ -419,82 +419,6 @@ public class MagenticOrchestrationTests
"final-answer synthesis must see what participants actually said");
}
[Fact]
public async Task Participant_Receives_Prior_Participant_Response_Not_InstructionAsync()
{
// Regression: each participant must see prior participants' *responses* (the running conversation),
// not their *instructions*. Previously the orchestrator broadcast the per-round instruction to every
// participant (untargeted fan-out) and never broadcast replies, so a later speaker received the earlier
// speaker's instruction and never its answer.
const string HealthInstruction = "HEALTH_CHECKER_INSTRUCTION_check_framework";
const string DatabaseInstruction = "DATABASE_CHECKER_INSTRUCTION_check_database";
const string HealthEchoPrefix = "HC_RESPONSE::";
const string DatabaseEchoPrefix = "DB_RESPONSE::";
List<ChatMessage> facts = CreatePlanResponse("Facts");
List<ChatMessage> plan = CreatePlanResponse("Plan");
List<ChatMessage> round1Ledger = CreateProgressLedgerResponse(
isRequestSatisfied: false,
isInLoop: false,
isProgressBeingMade: true,
nextSpeaker: "HealthChecker",
instructionOrQuestion: HealthInstruction);
List<ChatMessage> round2Ledger = CreateProgressLedgerResponse(
isRequestSatisfied: false,
isInLoop: false,
isProgressBeingMade: true,
nextSpeaker: "DatabaseChecker",
instructionOrQuestion: DatabaseInstruction);
List<ChatMessage> round3Ledger = CreateProgressLedgerResponse(
isRequestSatisfied: true,
isInLoop: false,
isProgressBeingMade: true,
nextSpeaker: "DatabaseChecker",
instructionOrQuestion: "Done");
List<ChatMessage> finalAnswer = CreateFinalAnswerResponse("All systems checked");
TestReplayAgent manager = new(
[facts, plan, round1Ledger, round2Ledger, round3Ledger, finalAnswer],
name: "Manager");
RecordingEchoAgent healthChecker = new(name: "HealthChecker", prefix: HealthEchoPrefix);
RecordingEchoAgent databaseChecker = new(name: "DatabaseChecker", prefix: DatabaseEchoPrefix);
Workflow workflow = new MagenticWorkflowBuilder(manager)
.AddParticipants(healthChecker, databaseChecker)
.RequirePlanSignoff(false)
.Build();
WorkflowRunResult runResult = await RunMagenticWorkflowAsync(
workflow,
[new ChatMessage(ChatRole.User, "Check system health")]);
runResult.Result.Should().NotBeNull();
runResult.Result![0].Text.Should().Contain("All systems checked");
// Each participant takes exactly one turn.
healthChecker.RecordedInputs.Should().ContainSingle();
databaseChecker.RecordedInputs.Should().ContainSingle();
// The first speaker receives its own instruction.
List<ChatMessage> healthInput = healthChecker.RecordedInputs[0];
healthInput.Should().Contain(m => m.Text.Contains(HealthInstruction), "the first speaker receives its own instruction");
// The second speaker must see the first speaker's RESPONSE (authored by HealthChecker, carrying the echo
// prefix that only the response — not the raw instruction — has), plus its own instruction.
List<ChatMessage> databaseInput = databaseChecker.RecordedInputs[0];
databaseInput.Should().Contain(
m => m.AuthorName == "HealthChecker" && m.Text.Contains(HealthEchoPrefix),
"the next speaker must receive the prior participant's response (the running conversation)");
databaseInput.Should().Contain(m => m.Text.Contains(DatabaseInstruction),
"the next speaker must receive its own instruction");
// The leaked-instruction bug: the second speaker must not receive HealthChecker's instruction as a
// bare message (it should only appear, if at all, embedded in HealthChecker's prefixed response).
databaseInput.Should().NotContain(
m => m.AuthorName != "HealthChecker" && m.Text.Trim() == HealthInstruction,
"the prior speaker's instruction must not leak into the next speaker's context as a standalone message");
}
[Fact]
public async Task PlanReview_Revised_Triggers_ReplanAsync()
{
@@ -1,37 +0,0 @@
// Copyright (c) Microsoft. All rights reserved.
using System.Collections.Generic;
using System.Linq;
using System.Runtime.CompilerServices;
using System.Threading;
using Microsoft.Extensions.AI;
namespace Microsoft.Agents.AI.Workflows.UnitTests;
/// <summary>
/// A <see cref="TestEchoAgent"/> that records the input messages it receives on each call.
/// Used by tests that need to assert what context a participant was actually handed - for example,
/// that a later speaker sees prior participants' <em>responses</em> (the running conversation) rather
/// than their <em>instructions</em>.
/// </summary>
internal sealed class RecordingEchoAgent(string? id = null, string? name = null, string? prefix = null)
: TestEchoAgent(id, name, prefix)
{
public List<List<ChatMessage>> RecordedInputs { get; } = [];
protected override async IAsyncEnumerable<AgentResponseUpdate> RunCoreStreamingAsync(
IEnumerable<ChatMessage> messages,
AgentSession? session = null,
AgentRunOptions? options = null,
[EnumeratorCancellation] CancellationToken cancellationToken = default)
{
// Materialize once so the deferred input is recorded and replayed identically.
List<ChatMessage> recorded = messages.ToList();
this.RecordedInputs.Add(recorded);
await foreach (AgentResponseUpdate update in base.RunCoreStreamingAsync(recorded, session, options, cancellationToken))
{
yield return update;
}
}
}
+1 -22
View File
@@ -7,26 +7,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [Unreleased]
## [1.8.1] - 2026-06-09
### Added
- **agent-framework-core**: Add MCP client OTel spans per GenAI semantic conventions ([#6349](https://github.com/microsoft/agent-framework/pull/6349))
- **agent-framework-core**: Add MCP long-running task support ([#6319](https://github.com/microsoft/agent-framework/pull/6319))
### Changed
- **agent-framework-claude**: Bump `claude-agent-sdk` to 0.2.87 ([#6248](https://github.com/microsoft/agent-framework/pull/6248))
- **agent-framework-core**: Document checkpoint storage security model and deserialization trust boundaries ([#6295](https://github.com/microsoft/agent-framework/pull/6295))
- **agent-framework-azurefunctions**: Document checkpoint storage security model and deserialization trust boundaries ([#6295](https://github.com/microsoft/agent-framework/pull/6295))
### Fixed
- **agent-framework-core**: Filter MCP tool kwargs to declared params via allowlist ([#6399](https://github.com/microsoft/agent-framework/pull/6399))
- **agent-framework-core**: Fix per-service-call history persistence with server-storing clients ([#6310](https://github.com/microsoft/agent-framework/pull/6310))
- **agent-framework-openai**: Use `getattr` for non-OpenAI provider response compatibility ([#6270](https://github.com/microsoft/agent-framework/pull/6270))
- **agent-framework-foundry-hosting**: Refactor workflow-as-agent pending request handling ([#6259](https://github.com/microsoft/agent-framework/pull/6259))
- **agent-framework-gemini**: Make Gemini honor declarative `outputSchema`, not just JSON mode ([#5893](https://github.com/microsoft/agent-framework/pull/5893))
- **agent-framework-mem0**: Isolate entity retrieval and correct `app_id` payload ([#6242](https://github.com/microsoft/agent-framework/pull/6242))
- **agent-framework-ag-ui**: Match AG-UI approval responses to requested arguments ([#6376](https://github.com/microsoft/agent-framework/pull/6376))
## [1.8.0] - 2026-06-04
### Added
@@ -1189,8 +1169,7 @@ Release candidate for **agent-framework-core** and **agent-framework-azure-ai**
For more information, see the [announcement blog post](https://devblogs.microsoft.com/foundry/introducing-microsoft-agent-framework-the-open-source-engine-for-agentic-ai-apps/).
[Unreleased]: https://github.com/microsoft/agent-framework/compare/python-1.8.1...HEAD
[1.8.1]: https://github.com/microsoft/agent-framework/compare/python-1.8.0...python-1.8.1
[Unreleased]: https://github.com/microsoft/agent-framework/compare/python-1.8.0...HEAD
[1.8.0]: https://github.com/microsoft/agent-framework/compare/python-1.7.0...python-1.8.0
[1.7.0]: https://github.com/microsoft/agent-framework/compare/python-1.6.0...python-1.7.0
[1.6.0]: https://github.com/microsoft/agent-framework/compare/python-1.5.0...python-1.6.0
@@ -9,7 +9,7 @@ from typing import Any, cast
from ag_ui.core import BaseEvent
from agent_framework import SupportsAgentRun
from ._agent_run import PendingApprovalEntry, run_agent_stream
from ._agent_run import run_agent_stream
class AgentConfig:
@@ -107,7 +107,7 @@ class AgentFrameworkAgent:
# Populated when approval requests are emitted; consumed when responses arrive.
# Prevents bypass, function name spoofing, and replay attacks.
# Bounded to prevent unbounded growth from abandoned approval requests.
self._pending_approvals: OrderedDict[str, PendingApprovalEntry] = OrderedDict()
self._pending_approvals: OrderedDict[str, str] = OrderedDict()
self._pending_approvals_max_size: int = 10_000
async def run(
@@ -8,7 +8,7 @@ import json
import logging
import uuid
from collections.abc import AsyncIterable, Awaitable
from typing import TYPE_CHECKING, Any, TypedDict, cast
from typing import TYPE_CHECKING, Any, cast
from ag_ui.core import (
BaseEvent,
@@ -56,7 +56,6 @@ from ._run_common import (
_stringify_tool_result, # type: ignore
)
from ._utils import (
canonical_function_arguments,
convert_agui_tools_to_agent_framework,
generate_event_id,
get_conversation_id_from_update,
@@ -408,33 +407,7 @@ def _make_approval_tool_result_events(resolved_approval_results: list[Content])
return events
class _PendingApproval(TypedDict):
"""Pending approval details for a requested function call."""
name: str
arguments: str | None
PendingApprovalEntry = _PendingApproval | str
def _make_pending_approval_entry(name: str, arguments: str | None) -> _PendingApproval:
return {"name": name, "arguments": arguments}
def _pending_approval_name(entry: PendingApprovalEntry) -> str | None:
if isinstance(entry, str):
return entry
return entry["name"]
def _pending_approval_arguments(entry: PendingApprovalEntry) -> str | None:
if isinstance(entry, str):
return None
return entry["arguments"]
def _evict_oldest_approvals(registry: dict[str, PendingApprovalEntry], max_size: int = 10_000) -> None:
def _evict_oldest_approvals(registry: dict[str, str], max_size: int = 10_000) -> None:
"""Evict the oldest entries from the pending-approvals registry (LRU).
Only effective when *registry* is an ``OrderedDict``; plain dicts are
@@ -454,7 +427,7 @@ async def _resolve_approval_responses(
tools: list[Any],
agent: SupportsAgentRun,
run_kwargs: dict[str, Any],
pending_approvals: dict[str, PendingApprovalEntry] | None = None,
pending_approvals: dict[str, str] | None = None,
thread_id: str = "",
) -> list[Content]:
"""Execute approved function calls and replace approval content with results.
@@ -507,8 +480,7 @@ async def _resolve_approval_responses(
invalid_ids.add(resp_id)
continue
pending_entry = pending_approvals[registry_key]
pending_name = _pending_approval_name(pending_entry)
pending_name = pending_approvals[registry_key]
if resp_name != pending_name:
logger.warning(
"Rejected approval response id=%s: function name mismatch (response=%s, pending=%s)",
@@ -519,16 +491,6 @@ async def _resolve_approval_responses(
invalid_ids.add(resp_id)
continue
pending_arguments = _pending_approval_arguments(pending_entry)
response_arguments = canonical_function_arguments(resp.function_call)
if pending_arguments is not None and response_arguments != pending_arguments:
logger.warning(
"Rejected approval response id=%s: function arguments mismatch",
resp_id,
)
invalid_ids.add(resp_id)
continue
# Valid — consume entry to prevent replay
del pending_approvals[registry_key]
if resp.approved:
@@ -752,7 +714,7 @@ async def run_agent_stream(
input_data: dict[str, Any],
agent: SupportsAgentRun,
config: AgentConfig,
pending_approvals: dict[str, PendingApprovalEntry] | None = None,
pending_approvals: dict[str, str] | None = None,
) -> AsyncGenerator[BaseEvent]:
"""Run agent and yield AG-UI events.
@@ -955,10 +917,7 @@ async def run_agent_stream(
# Register pending approval requests so we can validate responses later
if content_type == "function_approval_request" and pending_approvals is not None:
if content.id and content.function_call and content.function_call.name:
pending_approvals[f"{thread_id}:{content.id}"] = _make_pending_approval_entry(
content.function_call.name,
canonical_function_arguments(content.function_call),
)
pending_approvals[f"{thread_id}:{content.id}"] = content.function_call.name
# Evict oldest entries if the registry exceeds a safe bound (LRU)
_evict_oldest_approvals(pending_approvals, max_size=10_000)
else:
@@ -56,22 +56,6 @@ def safe_json_parse(value: Any) -> dict[str, Any] | None:
return None
def canonical_function_arguments(function_call: Any) -> str | None:
"""Return a stable representation of function-call arguments."""
if function_call is None:
return None
try:
parsed_arguments = function_call.parse_arguments()
except Exception:
parsed_arguments = getattr(function_call, "arguments", None)
if parsed_arguments is None:
parsed_arguments = {}
return json.dumps(make_json_safe(parsed_arguments), sort_keys=True, separators=(",", ":"))
def get_role_value(message: Any) -> str:
"""Extract role string from a message object.
@@ -35,7 +35,7 @@ from ._run_common import (
_extract_resume_payload,
_normalize_resume_interrupts,
)
from ._utils import canonical_function_arguments, generate_event_id, make_json_safe
from ._utils import generate_event_id, make_json_safe
logger = logging.getLogger(__name__)
@@ -324,29 +324,6 @@ def _coerce_response_for_request(request_event: Any, value: Any) -> Any | None:
return candidate
def _approval_response_matches_request(request_id: str, request_event: Any, response: Any) -> bool:
"""Check whether an approval response matches the pending approval request."""
request_data = getattr(request_event, "data", None)
if not isinstance(request_data, Content) or request_data.type != "function_approval_request":
return True
if not isinstance(response, Content) or response.type != "function_approval_response":
return False
if str(getattr(response, "id", "")) != request_id:
return False
request_call = getattr(request_data, "function_call", None)
response_call = getattr(response, "function_call", None)
if request_call is None or response_call is None:
return False
if getattr(response_call, "name", None) != getattr(request_call, "name", None):
return False
return canonical_function_arguments(response_call) == canonical_function_arguments(request_call)
def _single_pending_response_from_value(pending_events: dict[str, Any], value: Any) -> dict[str, Any]:
"""Map a scalar resume payload to the single pending request (if unambiguous)."""
if value is None or len(pending_events) != 1:
@@ -366,13 +343,6 @@ def _single_pending_response_from_value(pending_events: dict[str, Any], value: A
)
return {}
if not _approval_response_matches_request(str(request_id), request_event, coerced_value):
logger.info(
"Ignoring pending request response for request_id=%s: approval response does not match pending request",
request_id,
)
return {}
return {str(request_id): coerced_value}
@@ -402,12 +372,6 @@ def _coerce_responses_for_pending_requests(
_response_type_name(request_event),
)
continue
if not _approval_response_matches_request(request_key, request_event, coerced_value):
logger.info(
"Ignoring resume response for request_id=%s: approval response does not match pending request",
request_key,
)
continue
normalized[request_key] = coerced_value
return normalized
+2 -2
View File
@@ -1,6 +1,6 @@
[project]
name = "agent-framework-ag-ui"
version = "1.0.0rc4"
version = "1.0.0rc3"
description = "AG-UI protocol integration for Agent Framework"
readme = "README.md"
license-files = ["LICENSE"]
@@ -22,7 +22,7 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core>=1.8.1,<2",
"agent-framework-core>=1.6.0,<2",
"ag-ui-protocol>=0.1.16,<0.2",
"fastapi>=0.115.0,<0.133.1",
"uvicorn[standard]>=0.30.0,<1"
@@ -1407,92 +1407,6 @@ async def test_fabricated_rejection_without_pending_approval_is_blocked(streamin
assert False, "Fabricated rejection response leaked as function_result into LLM messages"
async def test_approval_argument_mismatch_is_blocked(streaming_chat_client_stub):
"""An approval response must not execute changed arguments for the pending call."""
from agent_framework import tool
from agent_framework.ag_ui import AgentFrameworkAgent
executed_args: list[dict[str, Any]] = []
@tool(
name="update_record",
description="Update a record",
approval_mode="always_require",
)
def update_record(record_id: str, value: str) -> str:
executed_args.append({"record_id": record_id, "value": value})
return f"updated {record_id} to {value}"
async def stream_fn_approval(
messages: MutableSequence[Message], options: ChatOptions, **kwargs: Any
) -> AsyncIterator[ChatResponseUpdate]:
yield ChatResponseUpdate(
contents=[
Content.from_function_call(
name="update_record",
call_id="call_update_001",
arguments={"record_id": "alpha", "value": "approved"},
)
]
)
wrapper = AgentFrameworkAgent(
agent=Agent(
client=streaming_chat_client_stub(stream_fn_approval),
name="test_agent",
instructions="Test",
tools=[update_record],
)
)
thread_id = "thread-argument-mismatch-test"
events1: list[Any] = []
async for event in wrapper.run({"thread_id": thread_id, "messages": [{"role": "user", "content": "update"}]}):
events1.append(event)
assert any("call_update_001" in k for k in wrapper._pending_approvals)
async def stream_fn_post(
messages: MutableSequence[Message], options: ChatOptions, **kwargs: Any
) -> AsyncIterator[ChatResponseUpdate]:
yield ChatResponseUpdate(contents=[Content.from_text(text="Done")])
wrapper.agent = Agent(
client=streaming_chat_client_stub(stream_fn_post),
name="test_agent",
instructions="Test",
tools=[update_record],
)
turn2_input: dict[str, Any] = {
"thread_id": thread_id,
"messages": [
{
"role": "user",
"content": "approve",
"function_approvals": [
{
"id": "call_update_001",
"call_id": "call_update_001",
"name": "update_record",
"approved": True,
"arguments": {"record_id": "beta", "value": "changed"},
}
],
},
],
}
events2: list[Any] = []
async for event in wrapper.run(turn2_input):
events2.append(event)
assert executed_args == []
assert any("call_update_001" in k for k in wrapper._pending_approvals), (
"Pending approval should be preserved after argument mismatch for legitimate retry"
)
async def test_state_update_end_to_end_via_real_tool_invocation(streaming_chat_client_stub):
"""End-to-end coverage for issue #3167: a real ``@tool`` returning ``state_update`` must
emit a deterministic STATE_SNAPSHOT through the full pipeline.
@@ -1352,70 +1352,6 @@ async def test_workflow_run_approval_via_messages_approved() -> None:
assert not resumed_finished.get("interrupt")
async def test_workflow_run_approval_argument_mismatch_keeps_interrupt_pending() -> None:
"""Workflow approval responses must not resume with changed function arguments."""
handled_responses: list[dict[str, Any]] = []
class ApprovalExecutor(Executor):
def __init__(self) -> None:
super().__init__(id="approval_executor")
@handler
async def start(self, message: Any, ctx: WorkflowContext) -> None:
del message
function_call = Content.from_function_call(
call_id="refund-call",
name="submit_refund",
arguments={"order_id": "12345", "amount": "$89.99"},
)
approval_request = Content.from_function_approval_request(id="approval-1", function_call=function_call)
await ctx.request_info(approval_request, Content, request_id="approval-1")
@response_handler
async def handle_approval(self, original_request: Content, response: Content, ctx: WorkflowContext) -> None:
del original_request
if response.function_call is not None:
handled_responses.append(response.function_call.parse_arguments() or {})
await ctx.yield_output("handled")
workflow = WorkflowBuilder(start_executor=ApprovalExecutor()).build()
first_events = [
event async for event in run_workflow_stream({"messages": [{"role": "user", "content": "go"}]}, workflow)
]
first_finished = [event for event in first_events if event.type == "RUN_FINISHED"][0].model_dump()
interrupt_payload = cast(list[dict[str, Any]], first_finished.get("interrupt"))
assert isinstance(interrupt_payload, list) and len(interrupt_payload) == 1
resumed_events = [
event
async for event in run_workflow_stream(
{
"messages": [
{
"role": "user",
"content": "",
"function_approvals": [
{
"approved": True,
"id": "approval-1",
"call_id": "refund-call",
"name": "submit_refund",
"arguments": {"order_id": "99999", "amount": "$1000.00"},
}
],
}
],
},
workflow,
)
]
assert handled_responses == []
resumed_finished = [event for event in resumed_events if event.type == "RUN_FINISHED"][0].model_dump()
assert resumed_finished.get("interrupt")
async def test_workflow_run_approval_via_messages_denied() -> None:
"""Denied approval response sent via messages (function_approvals) should satisfy the pending request."""
@@ -14,24 +14,6 @@ This module adds:
- reconstruct_to_type: for HITL responses where external data (without type markers)
needs to be reconstructed to a known type
- resolve_type: resolves 'module:class' type keys to Python types
Security Model
--------------
The underlying Azure Durable Functions storage (Azure Storage account) is the
trusted persistence layer for serialized checkpoint data. The
``RestrictedUnpickler`` in the core encoding module provides defense-in-depth
type filtering, but checkpoint storage itself must be properly access-controlled:
- Ensure the Azure Storage account used by Durable Functions is not publicly
writable and uses appropriate RBAC / shared-access policies.
- Never route untrusted user input directly into ``deserialize_value`` without
first calling :func:`strip_pickle_markers` to neutralize injection of
pickle markers into the data path.
- Configure your checkpoint storage with ``allowed_checkpoint_types`` (or call
``decode_checkpoint_value(..., allowed_types=...)`` directly) to restrict the set of types that can be deserialized.
See :mod:`agent_framework._workflows._checkpoint_encoding` for the full
security model documentation.
"""
from __future__ import annotations
@@ -4,7 +4,7 @@ description = "Azure Functions integration for Microsoft Agent Framework."
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.0.0b260609"
version = "1.0.0b260604"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -22,7 +22,7 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core>=1.8.1,<2",
"agent-framework-core>=1.8.0,<2",
"agent-framework-durabletask>=1.0.0b260604,<2",
"azure-functions>=1.24.0,<2",
"azure-functions-durable>=1.3.1,<2",
@@ -221,31 +221,9 @@ class ClaudeAgentOptions(TypedDict, total=False):
thinking: ThinkingConfig
"""Extended thinking configuration (adaptive, enabled, or disabled)."""
effort: Literal["low", "medium", "high", "xhigh", "max"]
effort: Literal["low", "medium", "high", "max"]
"""Effort level for thinking depth."""
skills: list[str] | Literal["all"]
"""Skills to enable for the main session. Use ``"all"`` for every discovered skill,
a list of named skills, or ``[]`` to suppress all skills."""
session_id: str
"""Use a specific session ID (must be a valid UUID) instead of auto-generated."""
task_budget: dict[str, int]
"""API-side task budget in tokens for pacing tool use."""
include_hook_events: bool
"""When True, hook lifecycle events are emitted in the message stream."""
strict_mcp_config: bool
"""When True, only use MCP servers passed via ``mcp_servers``, ignoring all others."""
continue_conversation: bool
"""Continue the most recent conversation instead of starting a new one."""
fork_session: bool
"""When True, resumed sessions fork to a new session ID."""
on_function_approval: FunctionApprovalCallback
"""Approval callback for ``FunctionTool`` instances declared with
``approval_mode="always_require"``. The callback is awaited (sync or async)
+3 -3
View File
@@ -4,7 +4,7 @@ description = "Claude Agent SDK integration for Microsoft Agent Framework."
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.0.0b260609"
version = "1.0.0b260521"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -23,8 +23,8 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core>=1.8.1,<2",
"claude-agent-sdk>=0.1.36,<0.3",
"agent-framework-core>=1.6.0,<2",
"claude-agent-sdk>=0.1.36,<0.1.49",
]
[tool.uv]
-2
View File
@@ -80,8 +80,6 @@ agent_framework/
- **`MCPTool`** - Base wrapper that owns the MCP `ClientSession` and exposes the remote server's tools as `FunctionTool`s.
- **`MCPStdioTool`** / **`MCPStreamableHTTPTool`** / **`MCPWebsocketTool`** - Transport-specific subclasses.
- **Argument allowlist (`_prepare_call_kwargs`)** - Before each `tools/call`, kwargs are filtered to an **allowlist** built from the tool's declared parameters (`inputSchema.properties`) plus any user-configured extras. Framework runtime kwargs injected through the function-invocation pipeline (e.g. `thread`, `conversation_id`, `chat_options`, `options`, `response_format`) are stripped by default rather than forwarded. A tool that declares no usable `properties` (including schemas with `additionalProperties: true`) forwards only the configured extras. The `_MCP_FRAMEWORK_DENYLIST` is a safety net for framework-named params a server *declares* in its schema (those are dropped); names explicitly opted in via `additional_tool_argument_names` always win. The reserved `_meta` key is extracted as MCP request metadata, never forwarded as an argument.
- **`additional_tool_argument_names`** (constructor arg on all `MCPTool` subclasses) - Opt extra argument names back into the allowlist. Accepts a `Sequence[str]` (applied to every tool) or a `Mapping[str, Sequence[str]]` keyed by **remote tool name**, where the reserved key `"*"` denotes global extras. It is configured only in user code at construction; there is **no per-call/runtime override**, so a model-issued tool call cannot change which names pass through. To use a server that accepts `additionalProperties: true`, list the extra names here and then either (1) manually extend that tool's `inputSchema` (via the `.functions` list after connecting) so the model is prompted to supply them, or (2) supply the values yourself via `function_invocation_kwargs`. If a name is supplied by both the model and `function_invocation_kwargs`, the model-supplied value wins.
- **`MCPTaskOptions`** (experimental, `MCP_LONG_RUNNING_TASKS` feature, **frozen**) - Per-tool-instance options controlling the SEP-2663 long-running task lifecycle. When the server advertises a tool with `execution.taskSupport == "required"`, `MCPTool.call_tool` transparently routes through `call_tool_as_task`, which sends an augmented `tools/call`, polls `tasks/get` until terminal, and reinterprets `tasks/result` as a normal `CallToolResult`. Instances are immutable; replace via `MCPTool.task_options = MCPTaskOptions(...)`. Fields:
- `default_ttl: timedelta | None` — forwarded to the server as `params.task.ttl` (milliseconds). When `None`, the server's default applies.
- `cancel_remote_task_on_local_cancellation: bool = True` — only gates the `CancelledError` path. Abandonment paths (see below) always cancel.
+29 -80
View File
@@ -92,16 +92,12 @@ OptionsCoT = TypeVar(
def _merge_options(base: dict[str, Any], override: dict[str, Any]) -> dict[str, Any]:
"""Merge two options dicts, with override values taking precedence.
``None`` is treated as "unset": ``None`` overrides are skipped so they don't clobber a base
value, and the merged result is stripped of any remaining ``None`` values in a final pass so
unset options are never forwarded (e.g. an unset ``store`` is left for the service to default).
Args:
base: The base options dict.
override: The override options dict (values take precedence).
Returns:
A new merged options dict containing no ``None`` values.
A new merged options dict.
"""
result = dict(base)
@@ -127,7 +123,7 @@ def _merge_options(base: dict[str, Any], override: dict[str, Any]) -> dict[str,
result["instructions"] = f"{result['instructions']}\n{value}"
else:
result[key] = value
return {key: value for key, value in result.items() if value is not None}
return result
def _sanitize_agent_name(agent_name: str | None) -> str | None:
@@ -464,9 +460,6 @@ class BaseAgent(SerializationMixin):
if provider_session is None and self.context_providers:
provider_session = AgentSession()
# When per-service-call persistence is enabled, the per-service-call middleware owns
# HistoryProvider persistence (in both the local and service-managed cases), so skip
# them on the once-per-run path to avoid double persistence.
per_service_call_history_required = self.require_per_service_call_history_persistence and any(
isinstance(provider, HistoryProvider) for provider in self.context_providers
)
@@ -693,16 +686,11 @@ class RawAgent(BaseAgent, Generic[OptionsCoT]): # type: ignore[misc]
description: A brief description of the agent's purpose.
context_providers: Context providers to include during agent invocation.
middleware: List of middleware to intercept agent and function invocations.
require_per_service_call_history_persistence: When True (and a HistoryProvider is
present), the provider always persists history via per-service-call middleware,
regardless of whether the client stores history server-side. If the client does
not store history, the middleware also loads providers around each model call and
drives the function loop with a local conversation; if it does, loading is skipped
(the service-managed conversation is the source of truth) and the middleware only
persists. A warning is logged for providers with ``load_messages=True`` when
loading is skipped because service-side storage is active. When no HistoryProvider
is present, this flag has no effect (no middleware is installed and nothing is
persisted).
require_per_service_call_history_persistence: When True, history providers are invoked
around each model call instead of once per ``run()`` when the service
is not already storing history. If service-side storage is active for
the run, the agent skips local history providers and relies on the
service-managed conversation instead.
default_options: A TypedDict containing chat options. When using a typed agent like
``Agent[OpenAIChatOptions]``, this enables IDE autocomplete for
provider-specific options including temperature, max_tokens, model,
@@ -803,20 +791,22 @@ class RawAgent(BaseAgent, Generic[OptionsCoT]): # type: ignore[misc]
self,
*,
session: AgentSession | None,
conversation_id: str | None,
options: Mapping[str, Any] | None,
service_stores_history: bool,
) -> list[HistoryProvider]:
history_providers = self._get_history_providers()
if not self.require_per_service_call_history_persistence or not history_providers:
return []
# A live service-managed session id takes precedence over the resolved conversation id.
if session and session.service_session_id:
conversation_id = session.service_session_id
# Without service-side storage the middleware persists locally and drives the function
# loop with a local sentinel, which cannot be reconciled with an existing service-managed
# conversation. When the service stores history, an existing conversation id is expected.
if conversation_id is not None and not service_stores_history:
conversation_id = (
session.service_session_id
if session and session.service_session_id
else cast(str | None, (options or {}).get("conversation_id") or self.default_options.get("conversation_id"))
)
if service_stores_history:
return []
if conversation_id is not None:
raise AgentInvalidRequestException(
"require_per_service_call_history_persistence cannot be used "
"with an existing service-managed conversation."
@@ -1177,34 +1167,18 @@ class RawAgent(BaseAgent, Generic[OptionsCoT]): # type: ignore[misc]
input_messages = normalize_messages(messages)
# Combine agent-level defaults with runtime options up front so the decisions below read
# `store` from a single place rather than introspecting both dicts. _merge_options applies
# the same precedence used for the actual client call (runtime wins; unset/None falls back
# to the agent default).
effective_options = _merge_options(self.default_options, opts)
# `store` in runtime or agent options takes precedence over the client's default
# storage behavior. An explicit `store=False` forces local (in-memory) history
# injection even when the client stores server-side by default; an explicit
# `store=True` forces service-side storage. A `store=None`/unset value means the
# service falls back to its own default.
explicit_store = effective_options.get("store")
# Internal behavior hint: will the service own history for this run? Only when the
# user left `store` unset do we fall back to the client's STORES_BY_DEFAULT.
service_stores_history = (
explicit_store if explicit_store is not None else getattr(self.client, "STORES_BY_DEFAULT", False)
)
# Resolve conversation_id from the same combined view so an agent-level default is honored
# when the runtime omits it (a live session id still takes precedence below).
effective_conversation_id = effective_options.get("conversation_id")
# `store` in runtime or agent options takes precedence over client-level storage
# indicators. An explicit `store=False` forces local (in-memory) history injection,
# even if the client is configured to use service-side storage by default.
store_ = opts.get("store", self.default_options.get("store", getattr(self.client, "STORES_BY_DEFAULT", False)))
# Auto-inject InMemoryHistoryProvider when session is provided, no context providers
# registered, and no service-side storage indicators
if (
session is not None
and not self.context_providers
and not session.service_session_id
and not effective_conversation_id
and not service_stores_history
and not opts.get("conversation_id")
and not store_
):
self.context_providers.append(InMemoryHistoryProvider())
@@ -1214,30 +1188,10 @@ class RawAgent(BaseAgent, Generic[OptionsCoT]): # type: ignore[misc]
per_service_call_history_providers = self._resolve_per_service_call_history_providers(
session=active_session,
conversation_id=effective_conversation_id,
service_stores_history=service_stores_history,
options=opts,
service_stores_history=bool(store_),
)
# When require_per_service_call_history_persistence is set together with a
# HistoryProvider, the per-service-call middleware (installed below) always persists
# the provider. ``service_stores_history`` only selects how the middleware behaves:
# - service does not store: the middleware also loads providers and drives the function
# loop with a local sentinel conversation id, or
# - service stores: the middleware skips loading (the service owns history) and simply
# persists each service call while the real conversation id flows through.
# In the service-managed case loading is skipped, so warn for providers that expect to load.
history_providers = self._get_history_providers()
if self.require_per_service_call_history_persistence and history_providers and service_stores_history:
for provider in history_providers:
if provider.load_messages:
logger.warning(
"HistoryProvider '%s' has load_messages=True but the chat client stores history "
"server-side; skipping local history load and relying on the service-managed "
"conversation. Set store=False to load from the provider, or load_messages=False "
"to silence this warning.",
provider.source_id,
)
session_context, chat_options = await self._prepare_session_and_messages(
session=active_session,
input_messages=input_messages,
@@ -1311,8 +1265,8 @@ class RawAgent(BaseAgent, Generic[OptionsCoT]): # type: ignore[misc]
}
if model is not None:
run_opts["model"] = model
# _merge_options strips unset (None) options, so e.g. an unset `store` is not forwarded
# and the service decides its own default.
# Remove None values and merge with chat_options
run_opts = {k: v for k, v in run_opts.items() if v is not None}
co = _merge_options(chat_options, run_opts)
# Build session_messages from session context: context messages + input messages
@@ -1326,7 +1280,6 @@ class RawAgent(BaseAgent, Generic[OptionsCoT]): # type: ignore[misc]
agent=self,
session=active_session,
providers=per_service_call_history_providers,
service_stores_history=service_stores_history,
)
existing_middleware = effective_client_kwargs.get("middleware")
if isinstance(existing_middleware, Sequence) and not isinstance(existing_middleware, (str, bytes)):
@@ -1366,7 +1319,7 @@ class RawAgent(BaseAgent, Generic[OptionsCoT]): # type: ignore[misc]
"input_messages": input_messages,
"session_messages": session_messages,
"agent_name": agent_name,
"suppress_response_id": bool(per_service_call_history_providers) and not service_stores_history,
"suppress_response_id": bool(per_service_call_history_providers),
"chat_options": co,
"compaction_strategy": compaction_strategy or self.compaction_strategy,
"tokenizer": tokenizer or self.tokenizer,
@@ -1460,15 +1413,11 @@ class RawAgent(BaseAgent, Generic[OptionsCoT]): # type: ignore[misc]
options=options or {},
)
# When per-service-call persistence is enabled, the per-service-call middleware owns
# HistoryProvider loading (it loads locally when the service does not store history, or
# relies on the service when it does), so skip them on the once-per-run before_run path.
per_service_call_history_required = self.require_per_service_call_history_persistence and bool(
self._get_history_providers()
)
# Run before_run providers (forward order, skip HistoryProvider when per-service-call
# persistence owns loading)
# Run before_run providers (forward order, skip HistoryProvider when per-service-call persistence owns history)
for provider in self.context_providers:
if per_service_call_history_required and isinstance(provider, HistoryProvider):
continue
@@ -604,13 +604,10 @@ class BaseChatClient(SerializationMixin, ABC, Generic[OptionsCoT]):
and dict literals are accepted without specialized option typing.
context_providers: Context providers to include during agent invocation.
middleware: List of middleware to intercept agent and function invocations.
require_per_service_call_history_persistence: When enabled (and a HistoryProvider is
present), the provider always persists history after each model call. If the
client does not store history server-side, history providers are also loaded and
injected around each model call; if it does, provider loading is skipped and the
service-managed conversation is the source of truth (persistence still happens
after each model call). When no HistoryProvider is present, this flag has no
effect (no middleware is installed and nothing is persisted).
require_per_service_call_history_persistence: Whether to require per-service-call
chat history persistence. When enabled, history providers are invoked around
each model call instead of once per ``run()`` when the service is not already
storing history.
function_invocation_configuration: Optional function invocation configuration override.
compaction_strategy: Optional agent-level compaction override. When omitted,
client-level compaction defaults remain in effect for each call.
+44 -160
View File
@@ -70,31 +70,6 @@ class MCPSpecificApproval(TypedDict, total=False):
_MCP_REMOTE_NAME_KEY = "_mcp_remote_name"
_MCP_NORMALIZED_NAME_KEY = "_mcp_normalized_name"
# Reserved key in an ``additional_tool_argument_names`` mapping that applies its
# values to every tool on the server rather than a single named tool.
_MCP_GLOBAL_EXTRA_ARGS_KEY = "*"
# Framework kwargs that flow through the function-invocation pipeline (via
# ``FunctionInvocationContext.kwargs``) but must never be forwarded to an MCP
# server: they are internal objects that the MCP SDK cannot serialize. They are
# dropped as a safety net when a tool declares one of them in its schema, unless
# the user explicitly opts the name back in via ``additional_tool_argument_names``
# (explicit extras always win over the denylist).
# - chat_options/tools/tool_choice/session/thread: framework runtime objects.
# - conversation_id: internal tracking ID used by services like Azure AI.
# - options: metadata/store used by AG-UI for Azure AI client requirements.
# - response_format: a Pydantic model class for structured output (not serializable).
# - _meta: reserved key extracted separately as MCP request metadata.
_MCP_FRAMEWORK_DENYLIST: frozenset[str] = frozenset({
"chat_options",
"tools",
"tool_choice",
"session",
"thread",
"conversation_id",
"options",
"response_format",
"_meta",
})
_mcp_call_headers: contextvars.ContextVar[dict[str, str]] = contextvars.ContextVar("_mcp_call_headers")
MCP_DEFAULT_TIMEOUT = 30
MCP_DEFAULT_SSE_READ_TIMEOUT = 60 * 5
@@ -160,34 +135,6 @@ def _build_prefixed_mcp_name(
return f"{normalized_prefix}_{trimmed_name}" if trimmed_name else normalized_prefix
def _normalize_additional_tool_argument_names(
additional_tool_argument_names: Sequence[str] | Mapping[str, Sequence[str]] | None,
) -> tuple[set[str], dict[str, set[str]]]:
"""Split user-supplied extra argument names into global and per-tool sets.
Accepts either a sequence (applied to every tool) or a mapping keyed by remote
tool name, where the reserved key ``"*"`` is treated as global. Mapping values
may be a sequence or a single string. Returns a
``(global_extras, per_tool_extras)`` tuple.
"""
if additional_tool_argument_names is None:
return set(), {}
if isinstance(additional_tool_argument_names, str):
return {additional_tool_argument_names}, {}
if isinstance(additional_tool_argument_names, Mapping):
global_extras: set[str] = set()
per_tool_extras: dict[str, set[str]] = {}
for tool_name, names in additional_tool_argument_names.items():
# Treat a bare string value as a single name rather than iterating its characters.
names_set = {names} if isinstance(names, str) else set(names)
if tool_name == _MCP_GLOBAL_EXTRA_ARGS_KEY:
global_extras.update(names_set)
else:
per_tool_extras[tool_name] = names_set
return global_extras, per_tool_extras
return set(additional_tool_argument_names), {}
def _inject_otel_into_mcp_meta(meta: dict[str, Any] | None = None) -> dict[str, Any] | None:
"""Inject OpenTelemetry trace context into MCP request _meta via the global propagator(s)."""
carrier: dict[str, str] = {}
@@ -347,7 +294,6 @@ class MCPTool:
client: SupportsChatGetResponse | None = None,
additional_properties: dict[str, Any] | None = None,
task_options: MCPTaskOptions | None = None,
additional_tool_argument_names: Sequence[str] | Mapping[str, Sequence[str]] | None = None,
) -> None:
"""Initialize the MCP Tool base.
@@ -382,10 +328,6 @@ class MCPTool:
task_options: Options controlling how long-running MCP tasks are driven for
tools that advertise ``execution.taskSupport == "required"``. When ``None``,
the defaults from :class:`MCPTaskOptions` are used.
additional_tool_argument_names: Extra argument names to forward to the MCP server
in addition to each tool's declared parameters. A ``Sequence[str]`` applies to
every tool; a ``Mapping[str, Sequence[str]]`` is keyed by remote tool name with
``"*"`` as a global key. See the transport subclasses for full details.
"""
self.name = name
self.description = description or ""
@@ -413,10 +355,6 @@ class MCPTool:
self._functions: list[FunctionTool] = []
self._tool_call_meta_by_name: dict[str, dict[str, Any]] = {}
self._tool_task_support_by_name: dict[str, str] = {}
self._tool_param_names_by_name: dict[str, set[str]] = {}
self._global_extra_arg_names, self._tool_extra_arg_names = _normalize_additional_tool_argument_names(
additional_tool_argument_names
)
self.is_connected: bool = False
self._tools_loaded: bool = False
self._prompts_loaded: bool = False
@@ -1291,7 +1229,6 @@ class MCPTool:
existing_names = {func.name for func in self._functions}
tool_call_meta_by_name: dict[str, dict[str, Any]] = {}
tool_task_support_by_name: dict[str, str] = {}
tool_param_names_by_name: dict[str, set[str]] = {}
params: types.PaginatedRequestParams | None = None
while True:
@@ -1334,24 +1271,6 @@ class MCPTool:
if task_support is not None:
tool_task_support_by_name[tool.name] = task_support
# Normalize inputSchema: ensure "properties" exists for object schemas.
# Some MCP servers (e.g. zero-argument tools) omit "properties",
# which causes OpenAI API to reject the schema with a 400 error.
# Guard against non-conforming MCP servers that send inputSchema=None
# despite the MCP spec typing it as dict[str, Any].
input_schema = dict(tool.inputSchema or {})
if input_schema.get("type") == "object" and "properties" not in input_schema:
input_schema["properties"] = {}
# Register declared param names before the existing-tool skip below so that
# reloads (e.g. notifications/tools/list_changed) preserve the allowlist for
# tools that are already loaded, consistent with tool_call_meta_by_name and
# tool_task_support_by_name above.
schema_properties = input_schema.get("properties")
tool_param_names_by_name[tool.name] = (
set(cast(dict[str, Any], schema_properties)) if isinstance(schema_properties, dict) else set()
)
normalized_name = _normalize_mcp_name(tool.name)
local_name = _build_prefixed_mcp_name(normalized_name, self.tool_name_prefix)
@@ -1360,6 +1279,14 @@ class MCPTool:
continue
approval_mode = self._determine_approval_mode(local_name, normalized_name, tool.name)
# Normalize inputSchema: ensure "properties" exists for object schemas.
# Some MCP servers (e.g. zero-argument tools) omit "properties",
# which causes OpenAI API to reject the schema with a 400 error.
# Guard against non-conforming MCP servers that send inputSchema=None
# despite the MCP spec typing it as dict[str, Any].
input_schema = dict(tool.inputSchema or {})
if input_schema.get("type") == "object" and "properties" not in input_schema:
input_schema["properties"] = {}
async def _call_tool_with_runtime_kwargs(
ctx: FunctionInvocationContext,
@@ -1393,7 +1320,6 @@ class MCPTool:
self._tool_call_meta_by_name = tool_call_meta_by_name
self._tool_task_support_by_name = tool_task_support_by_name
self._tool_param_names_by_name = tool_param_names_by_name
async def _close_on_owner(self) -> None:
# Cancel any pending reload tasks before tearing down the session.
@@ -1604,14 +1530,10 @@ class MCPTool:
raise ToolExecutionException(f"Failed to call tool '{tool_name}'.", inner_exception=ex) from ex
raise ToolExecutionException(f"Failed to call tool '{tool_name}' after retries.")
def _resolved_extra_args(self, tool_name: str) -> set[str]:
"""Return the user-configured extra argument names allowed for a tool."""
return self._global_extra_arg_names | self._tool_extra_arg_names.get(tool_name, set())
def _prepare_call_kwargs(
self, tool_name: str, kwargs: dict[str, Any]
) -> tuple[dict[str, Any], dict[str, Any] | None]:
"""Filter kwargs down to the tool's arguments and build the merged MCP request metadata."""
"""Filter framework-only kwargs and build the merged MCP request metadata."""
raw_user_meta: object | None = kwargs.get("_meta")
user_meta: dict[str, Any] | None = None
if raw_user_meta is not None and not isinstance(raw_user_meta, dict):
@@ -1624,28 +1546,27 @@ class MCPTool:
raise ToolExecutionException("MCP tool metadata provided via _meta must use string keys.")
user_meta[key] = value
# Allowlist: forward only the tool's declared parameters (from inputSchema.properties)
# plus any user-configured extra argument names. Everything else - notably the
# framework runtime kwargs injected through the function-invocation pipeline - is
# stripped so it is never forwarded to the MCP server. Tools that declare no usable
# properties forward only the user-configured extras.
#
# The extra names come exclusively from additional_tool_argument_names, which is set in
# user code at construction time; there is no per-call override, so a model-issued tool
# call cannot change which names are allowed through.
#
# The framework denylist acts as a safety net for keys a server *declares* in its
# schema that collide with internal, non-serializable framework objects (e.g. a tool
# that declares a parameter literally named "thread"): such declared-but-denylisted
# keys are dropped. Names the user explicitly opts in via additional_tool_argument_names
# always win. The reserved _meta key is handled separately above and never forwarded as
# an argument.
declared = self._tool_param_names_by_name.get(tool_name, set())
extras = self._resolved_extra_args(tool_name)
# Filter out framework kwargs that cannot be serialized by the MCP SDK.
# These are internal objects passed through the function invocation pipeline
# that should not be forwarded to external MCP servers.
# conversation_id is an internal tracking ID used by services like Azure AI.
# options contains metadata/store used by AG-UI for Azure AI client requirements.
# response_format is a Pydantic model class used for structured output (not serializable).
filtered_kwargs = {
k: v
for k, v in kwargs.items()
if k != "_meta" and (k in extras or (k in declared and k not in _MCP_FRAMEWORK_DENYLIST))
if k
not in {
"chat_options",
"tools",
"tool_choice",
"session",
"thread",
"conversation_id",
"options",
"response_format",
"_meta",
}
}
# Some MCP proxies require their tools/list metadata to be echoed on tools/call.
@@ -1722,7 +1643,9 @@ class MCPTool:
return parser(fallback_result)
if task_id is None:
raise ToolExecutionException(f"MCP server did not return a task_id or fallback result for '{tool_name}'.")
raise ToolExecutionException(
f"MCP server did not return a task_id or fallback result for '{tool_name}'."
)
# Track to completion: poll until terminal, then fetch payload. Never re-issue
# tools/call past this point; reconnect-and-retry only against the same task_id.
@@ -1842,7 +1765,9 @@ class MCPTool:
transient_codes: frozenset[int] = frozenset({int(httpx.codes.REQUEST_TIMEOUT)})
while True:
request = types.ClientRequest(types.GetTaskRequest(params=types.GetTaskRequestParams(taskId=task_id)))
request = types.ClientRequest(
types.GetTaskRequest(params=types.GetTaskRequestParams(taskId=task_id))
)
try:
# GetTaskResult.ttl is required-but-Optional in the SDK; coerce below.
lenient = await self._send_with_one_reconnect(
@@ -1850,7 +1775,9 @@ class MCPTool:
)
except McpError as ex:
if ex.error.code in transient_codes:
logger.debug("Transient %s on tasks/get for '%s'; will retry.", ex.error.code, task_id)
logger.debug(
"Transient %s on tasks/get for '%s'; will retry.", ex.error.code, task_id
)
await asyncio.sleep(_MCP_TASK_MIN_POLL_INTERVAL.total_seconds())
continue
# Hard server error mid-poll: task may still be running.
@@ -1979,7 +1906,9 @@ class MCPTool:
if not self._is_connection_lost(ex):
raise
if attempt < _MCP_RECONNECT_ATTEMPTS - 1:
logger.info("MCP connection lost during %s; reconnecting (task_id=%s).", operation, task_id)
logger.info(
"MCP connection lost during %s; reconnecting (task_id=%s).", operation, task_id
)
try:
await self.connect(reset=True)
except Exception as reconn_ex:
@@ -2038,7 +1967,9 @@ class MCPTool:
"""
from mcp import types
request = types.ClientRequest(types.CancelTaskRequest(params=types.CancelTaskRequestParams(taskId=task_id)))
request = types.ClientRequest(
types.CancelTaskRequest(params=types.CancelTaskRequestParams(taskId=task_id))
)
try:
await asyncio.wait_for(
self.session.send_request(request, types.CancelTaskResult), # type: ignore[union-attr]
@@ -2048,7 +1979,8 @@ class MCPTool:
raise
except asyncio.TimeoutError:
logger.warning(
"Best-effort tasks/cancel for '%s' timed out after %.1fs; remote task may still be running.",
"Best-effort tasks/cancel for '%s' timed out after %.1fs; "
"remote task may still be running.",
task_id,
_MCP_TASK_CANCEL_TIMEOUT.total_seconds(),
)
@@ -2221,7 +2153,6 @@ class MCPStdioTool(MCPTool):
client: SupportsChatGetResponse | None = None,
additional_properties: dict[str, Any] | None = None,
task_options: MCPTaskOptions | None = None,
additional_tool_argument_names: Sequence[str] | Mapping[str, Sequence[str]] | None = None,
**kwargs: Any,
) -> None:
"""Initialize the MCP stdio tool.
@@ -2268,20 +2199,6 @@ class MCPStdioTool(MCPTool):
client: The chat client to use for sampling.
task_options: Options for tools that advertise
``execution.taskSupport == "required"``. See :class:`MCPTaskOptions`.
additional_tool_argument_names: Extra argument names to forward to the MCP server in
addition to each tool's declared parameters (from its ``inputSchema.properties``).
By default only declared parameters are sent; framework runtime kwargs injected
through the function-invocation pipeline are stripped. Use this to opt specific
keys back in. Accepts either a ``Sequence[str]`` applied to every tool, or a
``Mapping[str, Sequence[str]]`` keyed by remote tool name where the reserved key
``"*"`` applies to every tool. This is configured only here in user code; there is
no per-call override, so a model-issued tool call cannot change which names pass
through. To use a server that accepts ``additionalProperties: true``, list the
extra names here and then either (1) manually extend that tool's ``inputSchema``
(via the ``.functions`` list after connecting) so the model is prompted to supply
them, or (2) supply the values yourself through ``function_invocation_kwargs``. If
a name is supplied via both the model and ``function_invocation_kwargs``, the
model-supplied value wins.
kwargs: Any extra arguments to pass to the stdio client.
"""
super().__init__(
@@ -2299,7 +2216,6 @@ class MCPStdioTool(MCPTool):
parse_prompt_results=parse_prompt_results,
request_timeout=request_timeout,
task_options=task_options,
additional_tool_argument_names=additional_tool_argument_names,
)
self.command = command
self.args = args or []
@@ -2379,7 +2295,6 @@ class MCPStreamableHTTPTool(MCPTool):
http_client: AsyncClient | None = None,
header_provider: Callable[[dict[str, Any]], dict[str, str]] | None = None,
task_options: MCPTaskOptions | None = None,
additional_tool_argument_names: Sequence[str] | Mapping[str, Sequence[str]] | None = None,
**kwargs: Any,
) -> None:
"""Initialize the MCP streamable HTTP tool.
@@ -2434,20 +2349,6 @@ class MCPStreamableHTTPTool(MCPTool):
agent middleware) without creating a separate ``httpx.AsyncClient``.
task_options: Options for tools that advertise
``execution.taskSupport == "required"``. See :class:`MCPTaskOptions`.
additional_tool_argument_names: Extra argument names to forward to the MCP server in
addition to each tool's declared parameters (from its ``inputSchema.properties``).
By default only declared parameters are sent; framework runtime kwargs injected
through the function-invocation pipeline are stripped. Use this to opt specific
keys back in. Accepts either a ``Sequence[str]`` applied to every tool, or a
``Mapping[str, Sequence[str]]`` keyed by remote tool name where the reserved key
``"*"`` applies to every tool. This is configured only here in user code; there is
no per-call override, so a model-issued tool call cannot change which names pass
through. To use a server that accepts ``additionalProperties: true``, list the
extra names here and then either (1) manually extend that tool's ``inputSchema``
(via the ``.functions`` list after connecting) so the model is prompted to supply
them, or (2) supply the values yourself through ``function_invocation_kwargs``. If
a name is supplied via both the model and ``function_invocation_kwargs``, the
model-supplied value wins.
kwargs: Additional keyword arguments (accepted for backward compatibility but not used).
"""
super().__init__(
@@ -2465,7 +2366,6 @@ class MCPStreamableHTTPTool(MCPTool):
parse_prompt_results=parse_prompt_results,
request_timeout=request_timeout,
task_options=task_options,
additional_tool_argument_names=additional_tool_argument_names,
)
self.url = url
self.terminate_on_close = terminate_on_close
@@ -2592,7 +2492,6 @@ class MCPWebsocketTool(MCPTool):
client: SupportsChatGetResponse | None = None,
additional_properties: dict[str, Any] | None = None,
task_options: MCPTaskOptions | None = None,
additional_tool_argument_names: Sequence[str] | Mapping[str, Sequence[str]] | None = None,
**kwargs: Any,
) -> None:
"""Initialize the MCP WebSocket tool.
@@ -2637,20 +2536,6 @@ class MCPWebsocketTool(MCPTool):
client: The chat client to use for sampling.
task_options: Options for tools that advertise
``execution.taskSupport == "required"``. See :class:`MCPTaskOptions`.
additional_tool_argument_names: Extra argument names to forward to the MCP server in
addition to each tool's declared parameters (from its ``inputSchema.properties``).
By default only declared parameters are sent; framework runtime kwargs injected
through the function-invocation pipeline are stripped. Use this to opt specific
keys back in. Accepts either a ``Sequence[str]`` applied to every tool, or a
``Mapping[str, Sequence[str]]`` keyed by remote tool name where the reserved key
``"*"`` applies to every tool. This is configured only here in user code; there is
no per-call override, so a model-issued tool call cannot change which names pass
through. To use a server that accepts ``additionalProperties: true``, list the
extra names here and then either (1) manually extend that tool's ``inputSchema``
(via the ``.functions`` list after connecting) so the model is prompted to supply
them, or (2) supply the values yourself through ``function_invocation_kwargs``. If
a name is supplied via both the model and ``function_invocation_kwargs``, the
model-supplied value wins.
kwargs: Any extra arguments to pass to the WebSocket client.
"""
super().__init__(
@@ -2668,7 +2553,6 @@ class MCPWebsocketTool(MCPTool):
parse_prompt_results=parse_prompt_results,
request_timeout=request_timeout,
task_options=task_options,
additional_tool_argument_names=additional_tool_argument_names,
)
self.url = url
self._client_kwargs = kwargs
@@ -16,7 +16,6 @@ from __future__ import annotations
import asyncio
import copy
import json
import logging
import threading
import uuid
import weakref
@@ -37,8 +36,6 @@ if TYPE_CHECKING:
from ._middleware import MiddlewareTypes
logger = logging.getLogger("agent_framework")
# Registry of known types for state deserialization
_STATE_TYPE_REGISTRY: dict[str, type] = {}
@@ -583,7 +580,6 @@ class PerServiceCallHistoryPersistingMiddleware(ChatMiddleware):
agent: SupportsAgentRun,
session: AgentSession,
providers: Sequence[HistoryProvider],
service_stores_history: bool = False,
) -> None:
"""Initialize the middleware.
@@ -591,16 +587,10 @@ class PerServiceCallHistoryPersistingMiddleware(ChatMiddleware):
agent: The agent that owns the history providers.
session: The active session for the current run.
providers: The history providers participating in per-service-call persistence.
service_stores_history: When True, the chat client stores history server-side. The
middleware then skips loading providers and leaves the real conversation id
untouched, persisting each service call without driving the function loop with a
local sentinel. When False, the middleware loads providers and uses a local
sentinel conversation id so the function loop runs without service-side storage.
"""
self._agent = agent
self._session = session
self._providers = list(providers)
self._service_stores_history = service_stores_history
async def _prepare_service_call_context(self, messages: Sequence[Message]) -> SessionContext:
"""Create a per-call SessionContext and load history providers into it."""
@@ -612,9 +602,6 @@ class PerServiceCallHistoryPersistingMiddleware(ChatMiddleware):
)
for source_id, source_messages in context_messages.items():
service_call_context.extend_messages(source_id, source_messages)
# When the service stores history, it owns loading; the providers are write-only sinks.
if self._service_stores_history:
return service_call_context
for provider in self._providers:
if not provider.load_messages:
continue
@@ -665,35 +652,17 @@ class PerServiceCallHistoryPersistingMiddleware(ChatMiddleware):
response: ChatResponse,
) -> ChatResponse:
"""Persist a model response and apply the local follow-up sentinel when needed."""
if (
not self._service_stores_history
and response.conversation_id is not None
and not is_local_history_conversation_id(response.conversation_id)
):
if response.conversation_id is not None and not is_local_history_conversation_id(response.conversation_id):
raise ChatClientInvalidResponseException(
"require_per_service_call_history_persistence cannot be used "
"when the chat client returns a real conversation_id."
)
# In storing mode the service is expected to echo a conversation id that the next run
# resumes from. If it comes back empty, the provider still captures this turn but there is
# no service id to load from next time, so cross-turn history can be lost silently. Warn
# every time so this uncommon, easy-to-miss failure mode cannot fail quietly.
if self._service_stores_history and response.conversation_id is None:
logger.warning(
"require_per_service_call_history_persistence is enabled with a chat client that "
"stores history server-side, but the client returned no conversation_id; cross-turn "
"history may not resume. Set store=False to load and resume from the HistoryProvider "
"instead."
)
await self._persist_service_call_response(
service_call_context=service_call_context,
response=response,
)
# The local sentinel only applies when the service does not store history; when it does,
# the real conversation id already drives function-loop continuation.
if not self._service_stores_history and _response_contains_follow_up_request(response):
if _response_contains_follow_up_request(response):
response.mark_internal_conversation_id()
response.conversation_id = LOCAL_HISTORY_CONVERSATION_ID
return response
@@ -712,12 +681,8 @@ class PerServiceCallHistoryPersistingMiddleware(ChatMiddleware):
result type for streaming or non-streaming execution.
"""
service_call_context = await self._prepare_service_call_context(context.messages)
# When the service stores history, leave the outgoing messages and the real conversation
# id untouched (pass-through); the middleware only persists. Otherwise reconstruct the
# outgoing messages from the loaded local history and strip the local sentinel.
if not self._service_stores_history:
context.messages = service_call_context.get_messages(include_input=True)
self._strip_local_conversation_id(context)
context.messages = service_call_context.get_messages(include_input=True)
self._strip_local_conversation_id(context)
await call_next()
@@ -13,35 +13,6 @@ during deserialization. The default built-in safe set covers common Python
value types (primitives, datetime, uuid, ...), all ``agent_framework`` internal
types, and all ``openai.types`` types. Callers can extend the set by passing
additional ``"module:qualname"`` strings.
Security Model
--------------
Checkpoint storage is treated as a **trusted data source**. The serialization
format uses Python's ``pickle`` module which can execute arbitrary code during
deserialization. The ``RestrictedUnpickler`` provides a defense-in-depth
allowlist that limits instantiable classes, but it is **not** a security
boundary — certain allowlisted builtins (e.g. ``getattr``) are required for
legitimate object reconstruction (enums, named tuples) and cannot be removed
without breaking compatibility.
Developers **must** ensure that:
1. The checkpoint storage backend (file system, Cosmos DB, Azure Blob, Durable
Functions storage) is access-controlled and not writable by untrusted
parties.
2. Data flowing into ``decode_checkpoint_value`` originates exclusively from
the application's own checkpoint storage — never from user-supplied HTTP
requests, message payloads, or other untrusted sources.
3. The ``allowed_types`` parameter is specified whenever possible to restrict
the set of reconstructible types to the minimum required by the application.
4. Never pass untrusted external input to ``decode_checkpoint_value``. If you
must accept external JSON that might contain checkpoint markers, sanitize it
first (for example, :func:`agent_framework_azurefunctions._serialization.strip_pickle_markers`).
The allowlist is a mitigation that reduces attack surface but does not
eliminate the inherent risks of deserializing untrusted pickle data. Treat
your checkpoint storage with the same access controls you would apply to
application secrets or database credentials.
"""
from __future__ import annotations
+1 -1
View File
@@ -4,7 +4,7 @@ description = "Microsoft Agent Framework for building AI Agents with Python. Thi
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.8.1"
version = "1.8.0"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
+2 -472
View File
@@ -3,7 +3,6 @@
import contextlib
import inspect
import json
import logging
from collections.abc import AsyncIterable, Awaitable, Callable, MutableSequence, Sequence
from typing import Any, cast
from unittest.mock import AsyncMock, MagicMock, patch
@@ -43,8 +42,6 @@ from agent_framework._mcp import MCPTool, _build_prefixed_mcp_name, _normalize_m
from agent_framework._middleware import FunctionInvocationContext
from agent_framework.exceptions import AgentInvalidRequestException, ChatClientInvalidResponseException
from .conftest import MockBaseChatClient
class _FixedTokenizer:
def __init__(self, token_count: int) -> None:
@@ -612,7 +609,6 @@ async def test_streaming_per_service_call_persistence_hides_response_id_from_aft
async def test_per_service_call_persistence_uses_real_service_storage_when_client_stores_by_default(
chat_client_base: SupportsChatGetResponse,
caplog: pytest.LogCaptureFixture,
) -> None:
provider = _RecordingHistoryProvider()
@@ -653,22 +649,15 @@ async def test_per_service_call_persistence_uses_real_service_storage_when_clien
require_per_service_call_history_persistence=True,
)
with caplog.at_level(logging.WARNING, logger="agent_framework"):
result = await agent.run("What's the weather in Seattle?", session=session)
result = await agent.run("What's the weather in Seattle?", session=session)
provider_state = session.state[provider.source_id]
assert result.text == "It is sunny in Seattle."
assert result.response_id == "resp_call_2"
assert chat_client_base.call_count == 2
# The service owns the conversation, so the provider never loads (issue #5798).
assert "get_call_count" not in provider_state
# Persistence is owned by the per-service-call middleware: it persists once per service call
# (issue #5798: the provider must never be silently bypassed when the service stores history).
# This run makes two service calls (function call + final answer), so it persists twice.
assert provider_state["save_call_count"] == 2
# load_messages=True while the service stores history surfaces a warning.
assert any("load_messages" in record.message for record in caplog.records)
assert "save_call_count" not in provider_state
assert session.service_session_id == "resp_service_managed"
@@ -2007,19 +1996,6 @@ def test_merge_options_none_values_ignored():
assert result["key2"] == "value2"
def test_merge_options_drops_none_base_values():
"""Test _merge_options strips None values so unset options are never forwarded."""
base = {"store": None, "temperature": 0.5}
override = {"top_p": 0.9}
result = _merge_options(base, override)
# An unset base value (e.g. store=None from default_options) must not survive the merge.
assert "store" not in result
assert result["temperature"] == 0.5
assert result["top_p"] == 0.9
def test_merge_options_runtime_model_overrides_default_model() -> None:
"""Test _merge_options lets a runtime model override a default model."""
result = _merge_options({"model": "default-model"}, {"model": "runtime-model"})
@@ -2682,449 +2658,3 @@ async def test_as_tool_raises_on_user_input_request(client: SupportsChatGetRespo
assert len(exc_info.value.contents) == 1
assert exc_info.value.contents[0].type == "oauth_consent_request"
assert exc_info.value.contents[0].consent_link == "https://login.microsoftonline.com/consent"
# region Per-service-call history persistence scenario matrix
#
# The driving field is ``require_per_service_call_history_persistence``. Every scenario runs a
# single agent run that makes **two service calls** -- a function call followed by a final
# completion -- so the *timing* of persistence is observable:
#
# * When the flag is ``True``, the per-service-call middleware persists the provider **after each
# service call**. So the function-call turn is already saved by the time the second (final)
# service call starts. This holds regardless of whether the chat client stores history
# server-side (the bug in issue #5798 was that a storing client silently bypassed persistence).
# * When the flag is ``False``, the provider persists **once, at the end of the run** -- nothing is
# saved between the two service calls.
#
# ``SpyChatClient.saves_before_call`` records ``provider.save_calls`` at the start of every service
# call, so ``[0, 1]`` means "the function-call turn was persisted before the final call" and
# ``[0, 0]`` means "no persistence happened mid-run". The client's ``store`` / ``STORES_BY_DEFAULT``
# only selects *how* the middleware behaves -- never *whether* the provider persists.
_PSC_SERVICE_CONVERSATION_ID = "svc-conversation"
_psc_stream_params = pytest.mark.parametrize("stream", [False, True], ids=["sync", "stream"])
@tool(name="lookup_weather", approval_mode="never_require")
def _psc_lookup_weather(location: str) -> str:
return f"Weather in {location}: sunny"
def _psc_function_call_script() -> list[tuple[str, ...]]:
"""A fresh function-call-then-final-completion script (the client mutates it)."""
return [
("call", "call_1", "lookup_weather", '{"location": "Seattle"}'),
("text", "It is sunny in Seattle."),
]
class _PscSpyHistoryProvider(HistoryProvider):
"""In-memory history provider that records load/save calls for assertions."""
def __init__(self, source_id: str = "spy_history", **kwargs: Any) -> None:
super().__init__(source_id, **kwargs)
self._messages: list[Message] = []
self.get_calls: int = 0
self.save_calls: int = 0
self.saved_batches: list[list[Message]] = []
async def get_messages(
self, session_id: str | None, *, state: dict[str, Any] | None = None, **kwargs: Any
) -> list[Message]:
self.get_calls += 1
return list(self._messages)
async def save_messages(
self,
session_id: str | None,
messages: Sequence[Message],
*,
state: dict[str, Any] | None = None,
**kwargs: Any,
) -> None:
self.save_calls += 1
self.saved_batches.append(list(messages))
self._messages.extend(messages)
@property
def stored_messages(self) -> list[Message]:
return list(self._messages)
class _PscSpyChatClient(MockBaseChatClient):
"""Chat client that scripts a function-call/final-completion sequence.
It records, at the start of each service call, how many provider saves have already happened
(``saves_before_call``), what messages it received, and what options it saw. When the effective
``store`` is truthy it returns a stable ``conversation_id`` to mimic a server-managed
conversation, so the framework propagates ``session.service_session_id``.
"""
def __init__(
self,
*,
provider: _PscSpyHistoryProvider,
stores_by_default: bool = False,
script: list[tuple[str, ...]] | None = None,
echo_conversation_id: bool = True,
**kwargs: Any,
) -> None:
super().__init__(**kwargs)
self.STORES_BY_DEFAULT = stores_by_default # type: ignore[attr-defined]
self._provider = provider
self._script = list(script) if script is not None else [("text", "ok")]
self._echo_conversation_id = echo_conversation_id
self.received_messages: list[list[Message]] = []
self.received_options: list[dict[str, Any]] = []
self.saves_before_call: list[int] = []
def _effective_store(self, options: dict[str, Any]) -> bool:
store = options.get("store")
if store is None:
return bool(self.STORES_BY_DEFAULT)
return bool(store)
def _next_contents(self) -> list[Content]:
turn = self._script.pop(0) if self._script else ("text", "ok")
if turn[0] == "call":
_, call_id, name, args = turn
return [Content.from_function_call(call_id=call_id, name=name, arguments=args)]
return [Content.from_text(turn[1])]
def _inner_get_response( # type: ignore[override]
self,
*,
messages: MutableSequence[Message],
stream: bool,
options: dict[str, Any],
**kwargs: Any,
) -> Awaitable[ChatResponse] | ResponseStream[ChatResponseUpdate, ChatResponse]:
self.received_messages.append(list(messages))
self.received_options.append(dict(options))
self.saves_before_call.append(self._provider.save_calls)
store_and_echo = self._effective_store(options) and self._echo_conversation_id
conv_id = _PSC_SERVICE_CONVERSATION_ID if store_and_echo else None
contents = self._next_contents()
if stream:
async def _stream() -> AsyncIterable[ChatResponseUpdate]:
self.call_count += 1
yield ChatResponseUpdate(
contents=contents,
role="assistant",
finish_reason="stop",
conversation_id=conv_id,
)
def _finalize(updates: Sequence[ChatResponseUpdate]) -> ChatResponse:
response = ChatResponse.from_updates(updates, output_format_type=options.get("response_format"))
if conv_id:
response.conversation_id = conv_id
return response
return ResponseStream(_stream(), finalizer=_finalize)
async def _get() -> ChatResponse:
self.call_count += 1
return ChatResponse(
messages=Message(role="assistant", contents=contents),
conversation_id=conv_id,
)
return _get()
def _psc_build_agent(
client: _PscSpyChatClient,
provider: _PscSpyHistoryProvider,
*,
require_per_service_call_history_persistence: bool,
default_options: dict[str, Any] | None = None,
) -> Agent:
kwargs: dict[str, Any] = {}
if default_options is not None:
kwargs["default_options"] = default_options
return Agent(
client=client,
tools=[_psc_lookup_weather],
context_providers=[provider],
require_per_service_call_history_persistence=require_per_service_call_history_persistence,
**kwargs,
)
async def _psc_run(agent: Agent, text: str, session: AgentSession, *, stream: bool) -> str:
if stream:
chunks: list[str] = []
async for update in agent.run(text, session=session, stream=True):
chunks.append(update.text or "")
return "".join(chunks)
result = await agent.run(text, session=session)
return result.text
# driver=True (the contract under test): persistence happens per service call
@_psc_stream_params
async def test_psc_flag_on_store_false_persists_after_each_service_call(stream: bool) -> None:
"""Mode A (flag on, service does not store): function-call turn is persisted before the final call."""
provider = _PscSpyHistoryProvider()
client = _PscSpyChatClient(provider=provider, stores_by_default=False, script=_psc_function_call_script())
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=True)
session = agent.create_session()
text = await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert text == "It is sunny in Seattle."
# Two service calls: function call, then final completion.
assert client.call_count == 2
# The contract: the function-call turn was persisted *before* the second service call started.
assert client.saves_before_call == [0, 1]
assert provider.save_calls == 2
# Mode A loads local history (the middleware injects it before each service call).
assert provider.get_calls >= 1
# No service-side storage, so no conversation id is propagated.
assert session.service_session_id is None
@_psc_stream_params
async def test_psc_flag_on_stores_by_default_persists_after_each_service_call(
stream: bool, caplog: pytest.LogCaptureFixture
) -> None:
"""Mode B (flag on, service stores by default): still persists per service call, but skips load (issue #5798)."""
provider = _PscSpyHistoryProvider() # load_messages=True by default
client = _PscSpyChatClient(provider=provider, stores_by_default=True, script=_psc_function_call_script())
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=True)
session = agent.create_session()
with caplog.at_level(logging.WARNING, logger="agent_framework"):
text = await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert text == "It is sunny in Seattle."
assert client.call_count == 2
# The invariant the bug violated: persistence still happens per service call when the service stores.
assert client.saves_before_call == [0, 1]
assert provider.save_calls == 2
# The service owns loading, so the provider is never asked to load.
assert provider.get_calls == 0
# A warning surfaces the bypassed load (load_messages=True).
assert any("load_messages" in record.message for record in caplog.records)
# The real service conversation id propagates to the session.
assert session.service_session_id == _PSC_SERVICE_CONVERSATION_ID
@_psc_stream_params
async def test_psc_flag_on_store_only_provider_no_load_no_warning(
stream: bool, caplog: pytest.LogCaptureFixture
) -> None:
"""Mode B with a store-only provider (load_messages=False): persists per call, no load, no warning."""
provider = _PscSpyHistoryProvider(load_messages=False)
client = _PscSpyChatClient(provider=provider, stores_by_default=True, script=_psc_function_call_script())
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=True)
session = agent.create_session()
with caplog.at_level(logging.WARNING, logger="agent_framework"):
await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert client.saves_before_call == [0, 1]
assert provider.save_calls == 2
assert provider.get_calls == 0
assert not any("load_messages" in record.message for record in caplog.records)
@_psc_stream_params
async def test_psc_flag_on_store_false_override_behaves_as_mode_a(stream: bool) -> None:
"""Flag on + storing client but store=False override: falls back to Mode A (local, per call)."""
provider = _PscSpyHistoryProvider()
client = _PscSpyChatClient(provider=provider, stores_by_default=True, script=_psc_function_call_script())
agent = _psc_build_agent(
client, provider, require_per_service_call_history_persistence=True, default_options={"store": False}
)
session = agent.create_session()
await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert client.saves_before_call == [0, 1]
assert provider.save_calls == 2
assert provider.get_calls >= 1
# store=False forces local handling, so no real service conversation id.
assert session.service_session_id is None
@_psc_stream_params
async def test_psc_flag_on_store_none_treated_as_absent(stream: bool, caplog: pytest.LogCaptureFixture) -> None:
"""Flag on + storing client + explicit store=None: None is "unset", so the storing default applies (Mode B)."""
provider = _PscSpyHistoryProvider()
client = _PscSpyChatClient(provider=provider, stores_by_default=True, script=_psc_function_call_script())
agent = _psc_build_agent(
client, provider, require_per_service_call_history_persistence=True, default_options={"store": None}
)
session = agent.create_session()
with caplog.at_level(logging.WARNING, logger="agent_framework"):
await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert client.saves_before_call == [0, 1]
assert provider.save_calls == 2
assert provider.get_calls == 0
assert session.service_session_id == _PSC_SERVICE_CONVERSATION_ID
assert any("load_messages" in record.message for record in caplog.records)
# store=None must not be forwarded to the client; the service decides its own default.
assert all("store" not in options for options in client.received_options)
@_psc_stream_params
async def test_psc_flag_on_respects_store_outputs_flag(stream: bool) -> None:
"""Flag on: the provider's store_inputs/store_outputs flags still apply per service call."""
provider = _PscSpyHistoryProvider(store_inputs=True, store_outputs=False)
client = _PscSpyChatClient(provider=provider, stores_by_default=True, script=_psc_function_call_script())
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=True)
session = agent.create_session()
await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert provider.save_calls == 2
# Outputs disabled, so no assistant/tool-call messages were stored, only user/tool inputs.
assert provider.stored_messages
assert all(message.role != "assistant" for message in provider.stored_messages)
# driver=False (control): persistence happens once, at the end of the run
@_psc_stream_params
async def test_psc_flag_off_store_false_persists_once_at_end(stream: bool) -> None:
"""Flag off + non-storing client: nothing is persisted mid-run; one save at the end."""
provider = _PscSpyHistoryProvider()
client = _PscSpyChatClient(provider=provider, stores_by_default=False, script=_psc_function_call_script())
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=False)
session = agent.create_session()
text = await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert text == "It is sunny in Seattle."
assert client.call_count == 2
# The control contract: no save happened between the function call and the final completion.
assert client.saves_before_call == [0, 0]
assert provider.save_calls == 1
@_psc_stream_params
async def test_psc_flag_off_stores_by_default_persists_once_at_end(stream: bool) -> None:
"""Flag off + storing client: once-per-run persistence, and the service conversation id propagates."""
provider = _PscSpyHistoryProvider()
client = _PscSpyChatClient(provider=provider, stores_by_default=True, script=_psc_function_call_script())
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=False)
session = agent.create_session()
await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert client.saves_before_call == [0, 0]
assert provider.save_calls == 1
assert session.service_session_id == _PSC_SERVICE_CONVERSATION_ID
@_psc_stream_params
async def test_psc_flag_on_storing_with_existing_conversation_id_does_not_raise(stream: bool) -> None:
"""Allow side of the guard: flag on + storing client + an existing conversation_id resumes (no raise).
The non-storing path raises on an existing service-managed conversation id, but with a storing
client the run must proceed and the service conversation id must propagate to the session.
"""
provider = _PscSpyHistoryProvider()
client = _PscSpyChatClient(provider=provider, stores_by_default=True, script=_psc_function_call_script())
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=True)
session = agent.create_session()
if stream:
chunks: list[str] = []
async for update in agent.run(
"What's the weather in Seattle?",
session=session,
stream=True,
options={"conversation_id": "existing_conversation"},
):
chunks.append(update.text or "")
text = "".join(chunks)
else:
result = await agent.run(
"What's the weather in Seattle?",
session=session,
options={"conversation_id": "existing_conversation"},
)
text = result.text
assert text == "It is sunny in Seattle."
# Persistence still happens per service call, and the real service id propagates to the session.
assert provider.save_calls == 2
assert provider.get_calls == 0
assert session.service_session_id == _PSC_SERVICE_CONVERSATION_ID
@_psc_stream_params
async def test_psc_flag_on_storing_two_runs_same_session(stream: bool) -> None:
"""Storing mode across two runs on one session: persistence keeps happening, id is stable, no load.
The second run exercises the precedence branch where the session already carries a
service_session_id, which must continue to skip provider loading and keep persisting.
"""
provider = _PscSpyHistoryProvider()
client = _PscSpyChatClient(provider=provider, stores_by_default=True, script=_psc_function_call_script())
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=True)
session = agent.create_session()
await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
assert provider.save_calls == 2
assert provider.get_calls == 0
first_run_service_id = session.service_session_id
assert first_run_service_id == _PSC_SERVICE_CONVERSATION_ID
# Reset the scripted client for a second run on the same session.
client._script = _psc_function_call_script()
client.call_count = 0
client.saves_before_call = []
await _psc_run(agent, "And in Portland?", session, stream=stream)
# Persistence keeps happening on the second run (two more saves), still per service call.
assert client.saves_before_call == [2, 3]
assert provider.save_calls == 4
# Loading stays skipped and the service conversation id stays stable across runs.
assert provider.get_calls == 0
assert session.service_session_id == first_run_service_id
@_psc_stream_params
async def test_psc_flag_on_storing_without_conversation_id_warns_every_call(
stream: bool, caplog: pytest.LogCaptureFixture
) -> None:
"""Storing mode but the client returns no conversation_id: warn on every service call.
Without an echoed conversation id the next run has nothing to resume from, so cross-turn
history can be lost silently. The warning fires per service call (no dedup) so the uncommon
failure mode cannot pass unnoticed.
"""
provider = _PscSpyHistoryProvider()
client = _PscSpyChatClient(
provider=provider,
stores_by_default=True,
script=_psc_function_call_script(),
echo_conversation_id=False,
)
agent = _psc_build_agent(client, provider, require_per_service_call_history_persistence=True)
session = agent.create_session()
with caplog.at_level(logging.WARNING, logger="agent_framework"):
await _psc_run(agent, "What's the weather in Seattle?", session, stream=stream)
# Persistence still happens, but no service id is captured to resume from.
assert provider.save_calls == 2
assert session.service_session_id is None
# Two service calls -> the warning is emitted twice (one per call, not deduped).
missing_id_warnings = [r for r in caplog.records if "returned no conversation_id" in r.message]
assert len(missing_id_warnings) == 2
-203
View File
@@ -30,7 +30,6 @@ from agent_framework._mcp import (
MCPTool,
_build_prefixed_mcp_name,
_get_input_model_from_mcp_prompt,
_normalize_additional_tool_argument_names,
_normalize_mcp_name,
_should_propagate_cancelled_error,
logger,
@@ -6058,205 +6057,3 @@ async def test_max_wait_interrupts_long_poll_sleep(monkeypatch: pytest.MonkeyPat
# endregion
# region additional_tool_argument_names / allowlist filtering
def test_normalize_additional_tool_argument_names_none() -> None:
global_extras, per_tool = _normalize_additional_tool_argument_names(None)
assert global_extras == set()
assert per_tool == {}
def test_normalize_additional_tool_argument_names_sequence() -> None:
global_extras, per_tool = _normalize_additional_tool_argument_names(["a", "b", "a"])
assert global_extras == {"a", "b"}
assert per_tool == {}
def test_normalize_additional_tool_argument_names_single_string() -> None:
# A bare string must be treated as a single name, not split into characters.
global_extras, per_tool = _normalize_additional_tool_argument_names("conversation_id")
assert global_extras == {"conversation_id"}
assert per_tool == {}
def test_normalize_additional_tool_argument_names_mapping_with_global_key() -> None:
global_extras, per_tool = _normalize_additional_tool_argument_names({
"*": ["g1"],
"tool_a": ["a1", "a2"],
"tool_b": ["b1"],
})
assert global_extras == {"g1"}
assert per_tool == {"tool_a": {"a1", "a2"}, "tool_b": {"b1"}}
def test_normalize_additional_tool_argument_names_mapping_with_string_values() -> None:
# A bare string mapping value is a single name, not an iterable of characters.
global_extras, per_tool = _normalize_additional_tool_argument_names({
"*": "conversation_id",
"tool_a": "custom",
})
assert global_extras == {"conversation_id"}
assert per_tool == {"tool_a": {"custom"}}
def test_prepare_call_kwargs_strips_undeclared_arguments() -> None:
server = MCPTool(name="test_server")
server._tool_param_names_by_name = {"test_tool": {"param"}}
filtered, meta = server._prepare_call_kwargs(
"test_tool",
{"param": "value", "conversation_id": "c", "thread": object(), "unexpected": 1},
)
assert filtered == {"param": "value"}
assert meta is None
def test_prepare_call_kwargs_global_extras_allowed() -> None:
server = MCPTool(name="test_server", additional_tool_argument_names=["conversation_id"])
server._tool_param_names_by_name = {"test_tool": {"param"}}
filtered, _ = server._prepare_call_kwargs(
"test_tool",
{"param": "value", "conversation_id": "c", "options": {}},
)
assert filtered == {"param": "value", "conversation_id": "c"}
def test_prepare_call_kwargs_per_tool_and_global_extras() -> None:
server = MCPTool(
name="test_server",
additional_tool_argument_names={"*": ["conversation_id"], "test_tool": ["custom"]},
)
server._tool_param_names_by_name = {"test_tool": {"param"}, "other_tool": {"x"}}
filtered, _ = server._prepare_call_kwargs(
"test_tool",
{"param": "v", "conversation_id": "c", "custom": "y", "thread": object()},
)
assert filtered == {"param": "v", "conversation_id": "c", "custom": "y"}
# The per-tool extra does not leak to other tools; the global one still applies.
filtered_other, _ = server._prepare_call_kwargs(
"other_tool",
{"x": 1, "conversation_id": "c", "custom": "y"},
)
assert filtered_other == {"x": 1, "conversation_id": "c"}
def test_prepare_call_kwargs_denylist_guards_server_declared_names() -> None:
# The denylist is a safety net for framework-named params a server *declares* in its
# schema: they are dropped so internal objects never leak. Names explicitly opted in
# via extras always win.
server = MCPTool(name="test_server", additional_tool_argument_names=["conversation_id"])
server._tool_param_names_by_name = {"test_tool": {"param", "thread"}}
filtered, _ = server._prepare_call_kwargs(
"test_tool",
{"param": "v", "thread": object(), "conversation_id": "c"},
)
# "thread" is declared by the schema but denylisted -> dropped; conversation_id opted in -> kept.
assert filtered == {"param": "v", "conversation_id": "c"}
def test_prepare_call_kwargs_extras_override_denylist() -> None:
# Opting a denylisted framework name back in via extras takes precedence over the
# denylist safety net. "thread" is on the framework denylist, but an explicit extra wins.
server = MCPTool(name="test_server", additional_tool_argument_names=["thread"])
server._tool_param_names_by_name = {"test_tool": {"param"}}
sentinel = object()
filtered, _ = server._prepare_call_kwargs(
"test_tool",
{"param": "v", "thread": sentinel, "conversation_id": "c"},
)
# "thread" opted in via extras -> kept despite the denylist; conversation_id is denylisted,
# not declared, and not opted in -> dropped.
assert filtered == {"param": "v", "thread": sentinel}
def test_prepare_call_kwargs_zero_arg_tool_passes_no_arguments() -> None:
server = MCPTool(name="test_server")
server._tool_param_names_by_name = {"test_tool": set()}
filtered, _ = server._prepare_call_kwargs(
"test_tool",
{"conversation_id": "c", "thread": object(), "stray": 1},
)
assert filtered == {}
def test_prepare_call_kwargs_unknown_tool_passes_only_global_extras() -> None:
server = MCPTool(name="test_server", additional_tool_argument_names=["conversation_id"])
# No entry in _tool_param_names_by_name for this tool name.
filtered, _ = server._prepare_call_kwargs(
"unknown_tool",
{"conversation_id": "c", "other": 1},
)
assert filtered == {"conversation_id": "c"}
def test_prepare_call_kwargs_extracts_meta() -> None:
server = MCPTool(name="test_server")
server._tool_param_names_by_name = {"test_tool": {"param"}}
filtered, meta = server._prepare_call_kwargs(
"test_tool",
{"param": "v", "_meta": {"trace": "abc"}},
)
assert filtered == {"param": "v"}
assert meta is not None
assert meta.get("trace") == "abc"
async def test_call_tool_forwards_only_declared_arguments() -> None:
"""End-to-end: framework runtime kwargs are stripped before reaching the server."""
class TestServer(MCPTool):
async def connect(self):
self.session = Mock(spec=ClientSession)
self.session.list_tools = AsyncMock(
return_value=types.ListToolsResult(
tools=[
types.Tool(
name="test_tool",
description="Test tool",
inputSchema={
"type": "object",
"properties": {"param": {"type": "string"}},
"required": ["param"],
},
)
]
)
)
self.session.call_tool = AsyncMock(
return_value=types.CallToolResult(content=[types.TextContent(type="text", text="ok")])
)
def get_mcp_client(self) -> _AsyncGeneratorContextManager[Any, None]:
return None
server = TestServer(name="test_server", additional_tool_argument_names=["conversation_id"])
async with server:
await server.load_tools()
session_mock = server.session
await server.call_tool(
"test_tool",
param="value",
conversation_id="c",
thread=object(),
response_format=object(),
)
session_mock.call_tool.assert_called_once()
_, call_kwargs = session_mock.call_tool.call_args
assert call_kwargs["arguments"] == {"param": "value", "conversation_id": "c"}
# endregion
+3 -3
View File
@@ -4,7 +4,7 @@ description = "Microsoft Foundry integrations for Microsoft Agent Framework."
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.8.1"
version = "1.8.0"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -23,8 +23,8 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core>=1.8.1,<2",
"agent-framework-openai>=1.8.1,<2",
"agent-framework-core>=1.8.0,<2",
"agent-framework-openai>=1.8.0,<2",
"azure-ai-inference>=1.0.0b9,<1.0.0b10",
"azure-ai-projects>=2.2.0,<3.0",
]
@@ -4,7 +4,7 @@ description = "Foundry Hosting integration for Microsoft Agent Framework."
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.0.0a260609"
version = "1.0.0a260604"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -23,7 +23,7 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core>=1.8.1,<2",
"agent-framework-core>=1.8.0,<2",
"azure-ai-agentserver-core>=2.0.0b3,<3",
"azure-ai-agentserver-responses>=1.0.0b7,<2",
"azure-ai-agentserver-invocations>=1.0.0b3,<2",
+2 -2
View File
@@ -4,7 +4,7 @@ description = "Google Gemini integration for Microsoft Agent Framework."
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.0.0a260609"
version = "1.0.0a260521"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -24,7 +24,7 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core>=1.8.1,<2.0",
"agent-framework-core>=1.6.0,<2.0",
"google-genai>=1.65.0,<2.0.0",
]
@@ -8,34 +8,29 @@ This module provides ``Mem0ContextProvider``, built on the new
from __future__ import annotations
import asyncio
import logging
import sys
from collections.abc import Awaitable
from contextlib import AbstractAsyncContextManager
from typing import TYPE_CHECKING, Any, ClassVar, TypeAlias, TypedDict
from typing import TYPE_CHECKING, Any, ClassVar
from agent_framework import Message
from agent_framework._sessions import AgentSession, ContextProvider, SessionContext
from mem0 import AsyncMemory, AsyncMemoryClient
if sys.version_info >= (3, 11):
from typing import Self # pragma: no cover
from typing import NotRequired, Self, TypedDict # pragma: no cover
else:
from typing_extensions import Self # pragma: no cover
from typing_extensions import NotRequired, Self, TypedDict # pragma: no cover
if TYPE_CHECKING:
from agent_framework._agents import SupportsAgentRun
logger = logging.getLogger(__name__)
MemoryRecord: TypeAlias = dict[str, object]
class _MemorySearchResponse_v1_1(TypedDict):
results: list[dict[str, Any]]
relations: NotRequired[list[dict[str, Any]]]
class SearchResults(TypedDict):
results: list[MemoryRecord]
SearchResponse: TypeAlias = list[MemoryRecord] | SearchResults
_MemorySearchResponse_v2 = list[dict[str, Any]]
class Mem0ContextProvider(ContextProvider):
@@ -111,85 +106,28 @@ class Mem0ContextProvider(ContextProvider):
if not input_text.strip():
return
# Query entity partitions independently to bypass strict logical AND limitations
# Mem0 OSS and Platform SDKs expose inconsistent search typings.
search_tasks: list[Awaitable[Any]] = []
filters = self._build_filters()
# 1. Query User partition independently
if self.user_id:
user_kwargs = self._build_search_kwargs(input_text, "user_id", self.user_id)
search_tasks.append(self.mem0_client.search(**user_kwargs)) # type: ignore[reportUnknownMemberType, reportUnknownArgumentType]
# AsyncMemory (OSS) expects user_id/agent_id/run_id as direct kwargs
# AsyncMemoryClient (Platform) expects them in a filters dict
search_kwargs: dict[str, Any] = {"query": input_text}
if isinstance(self.mem0_client, AsyncMemory):
search_kwargs.update(filters)
else:
search_kwargs["filters"] = filters
# 2. Query Agent partition independently
if self.agent_id:
agent_kwargs = self._build_search_kwargs(input_text, "agent_id", self.agent_id)
search_tasks.append(self.mem0_client.search(**agent_kwargs)) # type: ignore[reportUnknownMemberType, reportUnknownArgumentType]
search_response: _MemorySearchResponse_v1_1 | _MemorySearchResponse_v2 = await self.mem0_client.search( # type: ignore[misc]
**search_kwargs,
)
# Fall back to an app-scoped search when only application_id is configured
if not search_tasks and self.application_id:
app_kwargs: dict[str, Any] = {"query": input_text}
if isinstance(self.mem0_client, AsyncMemory):
app_kwargs["app_id"] = self.application_id
else:
app_kwargs["filters"] = {"app_id": self.application_id}
search_tasks.append(self.mem0_client.search(**app_kwargs)) # pyright: ignore[reportUnknownMemberType, reportUnknownArgumentType]
if not search_tasks:
return
if isinstance(search_response, list):
memories = search_response
elif isinstance(search_response, dict) and "results" in search_response:
memories = search_response["results"]
else:
memories = [search_response]
results: list[SearchResponse | BaseException] = await asyncio.gather(*search_tasks, return_exceptions=True)
# Merge and deduplicate results
memories: list[MemoryRecord] = []
seen_memory_ids: set[str] = set()
failed_tasks_count: int = 0
for search_response in results:
if isinstance(search_response, asyncio.CancelledError):
raise search_response
if isinstance(search_response, BaseException):
failed_tasks_count += 1
logger.error(
"Mem0 partition search task failed: %s",
search_response,
exc_info=(type(search_response), search_response, search_response.__traceback__),
)
continue
current_memories: list[MemoryRecord] = []
if isinstance(search_response, list):
current_memories = [mem for mem in search_response if isinstance(mem, dict)]
elif isinstance(search_response, dict):
results_field = search_response.get("results")
if isinstance(results_field, list):
current_memories = [
item
for item in results_field
if isinstance(item, dict) # pyright: ignore[reportUnknownVariableType]
]
else:
logger.warning(
"Unexpected Mem0 search response format: %s",
type(results_field).__name__,
)
for mem in current_memories:
mem_id = mem.get("id")
if mem_id is not None and not isinstance(mem_id, str):
mem_id = str(mem_id)
if mem_id is not None and mem_id in seen_memory_ids:
continue
if mem_id is not None:
seen_memory_ids.add(mem_id)
memories.append(mem)
if failed_tasks_count == len(search_tasks):
logger.error("All Mem0 retrieval tasks failed. Context provider is unable to verify memory state.")
line_separated_memories = "\n".join(str(memory.get("memory", "")) for memory in memories)
line_separated_memories = "\n".join(memory.get("memory", "") for memory in memories)
if line_separated_memories:
context.extend_messages(
self.source_id,
@@ -221,21 +159,12 @@ class Mem0ContextProvider(ContextProvider):
]
if messages:
add_kwargs: dict[str, Any] = {
"messages": messages,
"user_id": self.user_id,
"agent_id": self.agent_id,
}
# Inject the application scope using the matching signature format for each SDK variant
if isinstance(self.mem0_client, AsyncMemory):
if self.application_id:
add_kwargs["app_id"] = self.application_id
else:
if self.application_id:
add_kwargs["filters"] = {"app_id": self.application_id}
await self.mem0_client.add(**add_kwargs) # type: ignore[misc, call-arg]
await self.mem0_client.add( # type: ignore[misc]
messages=messages,
user_id=self.user_id,
agent_id=self.agent_id,
metadata={"application_id": self.application_id},
)
# -- Internal methods ------------------------------------------------------
@@ -244,21 +173,15 @@ class Mem0ContextProvider(ContextProvider):
if not self.agent_id and not self.user_id and not self.application_id:
raise ValueError("At least one of the filters: agent_id, user_id, or application_id is required.")
def _build_search_kwargs(self, input_text: str, entity_key: str, entity_value: str) -> dict[str, Any]:
"""Build search keyword arguments formatted for OSS vs Platform clients."""
filters: dict[str, Any] = {"query": input_text}
if isinstance(self.mem0_client, AsyncMemory):
# AsyncMemory (OSS) expects direct kwargs
filters[entity_key] = entity_value
if self.application_id:
filters["app_id"] = self.application_id
else:
# AsyncMemoryClient (Platform) expects a filters dict
filters["filters"] = {entity_key: entity_value}
if self.application_id:
filters["filters"]["app_id"] = self.application_id
def _build_filters(self) -> dict[str, Any]:
"""Build search filters from initialization parameters."""
filters: dict[str, Any] = {}
if self.user_id:
filters["user_id"] = self.user_id
if self.agent_id:
filters["agent_id"] = self.agent_id
if self.application_id:
filters["app_id"] = self.application_id
return filters
+2 -2
View File
@@ -4,7 +4,7 @@ description = "Mem0 integration for Microsoft Agent Framework."
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.0.0b260609"
version = "1.0.0b260521"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -23,7 +23,7 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core>=1.8.1,<2",
"agent-framework-core>=1.6.0,<2",
"mem0ai>=1.0.0,<2",
]
@@ -3,7 +3,7 @@
from __future__ import annotations
from unittest.mock import AsyncMock, MagicMock, patch
from unittest.mock import AsyncMock, patch
import pytest
from agent_framework import AgentResponse, Message
@@ -193,59 +193,39 @@ class TestBeforeRun:
assert call_kwargs["user_id"] == "u1"
assert "filters" not in call_kwargs
@pytest.mark.asyncio
async def test_oss_client_all_scoping_params_except_app_id(self, mock_oss_mem0_client: AsyncMock) -> None:
"""OSS client with all scoping parameters passes them as isolated concurrent kwargs."""
async def test_oss_client_all_scoping_params(self, mock_oss_mem0_client: AsyncMock) -> None:
"""OSS client with all scoping parameters passes them as direct kwargs."""
mock_oss_mem0_client.search.return_value = []
provider = Mem0ContextProvider(
source_id="mem0",
mem0_client=mock_oss_mem0_client,
user_id="u1",
agent_id="a1"
source_id="mem0", mem0_client=mock_oss_mem0_client, user_id="u1", agent_id="a1", application_id="app1"
)
mock_context = MagicMock(spec=SessionContext)
mock_msg = MagicMock()
mock_msg.text = "hello"
mock_context.input_messages = [mock_msg]
mock_context.response = None
session = AgentSession(session_id="test-session")
ctx = SessionContext(input_messages=[Message(role="user", contents=["Hello"])], session_id="s1")
await provider.before_run(
agent=MagicMock(), session=MagicMock(spec=AgentSession), context=mock_context, state={}
)
agent=None, session=session, context=ctx, state=session.state.setdefault(provider.source_id, {})
) # type: ignore[arg-type]
# Re-aligned assertion: We expect 2 separate concurrent calls instead of 1 combined call
assert mock_oss_mem0_client.search.call_count == 2
mock_oss_mem0_client.search.assert_any_call(query="hello", user_id="u1")
mock_oss_mem0_client.search.assert_any_call(query="hello", agent_id="a1")
call_kwargs = mock_oss_mem0_client.search.call_args.kwargs
assert call_kwargs["user_id"] == "u1"
assert call_kwargs["agent_id"] == "a1"
assert "filters" not in call_kwargs
@pytest.mark.asyncio
async def test_platform_client_passes_filters_dict_except_app_id(self, mock_mem0_client: AsyncMock) -> None:
"""Platform client passes scoping parameters concurrently inside the nested filters dictionary."""
async def test_platform_client_passes_filters_dict(self, mock_mem0_client: AsyncMock) -> None:
"""Platform AsyncMemoryClient should receive scoping params in a filters dict."""
mock_mem0_client.search.return_value = []
provider = Mem0ContextProvider(
source_id="mem0",
mem0_client=mock_mem0_client,
user_id="u1",
agent_id="a1",
)
mock_context = MagicMock(spec=SessionContext)
mock_msg = MagicMock()
mock_msg.text = "hello"
mock_context.input_messages = [mock_msg]
mock_context.response = None
provider = Mem0ContextProvider(source_id="mem0", mem0_client=mock_mem0_client, user_id="u1")
session = AgentSession(session_id="test-session")
ctx = SessionContext(input_messages=[Message(role="user", contents=["Hello"])], session_id="s1")
await provider.before_run(
agent=MagicMock(), session=MagicMock(spec=AgentSession), context=mock_context, state={}
)
agent=None, session=session, context=ctx, state=session.state.setdefault(provider.source_id, {})
) # type: ignore[arg-type]
# Re-aligned assertion: Platform client isolates filters per call to bypass AND limitations
assert mock_mem0_client.search.call_count == 2
mock_mem0_client.search.assert_any_call(query="hello", filters={"user_id": "u1"})
mock_mem0_client.search.assert_any_call(query="hello", filters={"agent_id": "a1"})
call_kwargs = mock_mem0_client.search.call_args.kwargs
assert call_kwargs["query"] == "Hello"
assert "filters" in call_kwargs
assert call_kwargs["filters"]["user_id"] == "u1"
# -- after_run tests -----------------------------------------------------------
@@ -338,8 +318,8 @@ class TestAfterRun:
with pytest.raises(ValueError, match="At least one of the filters"):
await provider.after_run(agent=None, session=session, context=ctx, state=session.state) # type: ignore[arg-type]
async def test_stores_with_application_id_filters(self, mock_mem0_client: AsyncMock) -> None:
"""application_id is passed in filters."""
async def test_stores_with_application_id_metadata(self, mock_mem0_client: AsyncMock) -> None:
"""application_id is passed in metadata."""
provider = Mem0ContextProvider(
source_id="mem0", mem0_client=mock_mem0_client, user_id="u1", application_id="app1"
)
@@ -351,7 +331,7 @@ class TestAfterRun:
agent=None, session=session, context=ctx, state=session.state.setdefault(provider.source_id, {})
) # type: ignore[arg-type]
assert mock_mem0_client.add.call_args.kwargs["filters"] == {"app_id": "app1"}
assert mock_mem0_client.add.call_args.kwargs["metadata"] == {"application_id": "app1"}
# -- _validate_filters tests --------------------------------------------------
@@ -378,20 +358,15 @@ class TestValidateFilters:
provider._validate_filters()
# -- _build_search_kwargs tests -----------------------------------------------------
# -- _build_filters tests -----------------------------------------------------
class TestBuildSearchKwargs:
"""Test _build_search_kwargs method."""
class TestBuildFilters:
"""Test _build_filters method."""
def test_user_id_only(self, mock_mem0_client: AsyncMock) -> None:
provider = Mem0ContextProvider(source_id="mem0", mem0_client=mock_mem0_client, user_id="u1")
# Pass the 3 required arguments
result = provider._build_search_kwargs("test query", "user_id", "u1")
# AsyncMock triggers the Platform client nested 'filters' structure
assert result == {"query": "test query", "filters": {"user_id": "u1"}}
assert provider._build_filters() == {"user_id": "u1"}
def test_all_params(self, mock_mem0_client: AsyncMock) -> None:
provider = Mem0ContextProvider(
@@ -401,66 +376,28 @@ class TestBuildSearchKwargs:
agent_id="a1",
application_id="app1",
)
# Test that app_id correctly merges with the isolated target entity
result = provider._build_search_kwargs("test query", "agent_id", "a1")
assert result == {
"query": "test query",
"filters": {
"agent_id": "a1",
"app_id": "app1",
},
assert provider._build_filters() == {
"user_id": "u1",
"agent_id": "a1",
"app_id": "app1",
}
def test_excludes_none_values(self, mock_mem0_client: AsyncMock) -> None:
provider = Mem0ContextProvider(source_id="mem0", mem0_client=mock_mem0_client, user_id="u1")
# application_id is None by default, it should not appear in the dictionary
result = provider._build_search_kwargs("test query", "user_id", "u1")
assert "app_id" not in result.get("filters", {})
filters = provider._build_filters()
assert "agent_id" not in filters
assert "run_id" not in filters
assert "app_id" not in filters
def test_no_run_id_in_search_filters(self, mock_mem0_client: AsyncMock) -> None:
"""run_id is excluded from search filters so memories work across sessions."""
provider = Mem0ContextProvider(source_id="mem0", mem0_client=mock_mem0_client, user_id="u1")
result = provider._build_search_kwargs("test query", "user_id", "u1")
assert "run_id" not in result.get("filters", {})
assert "run_id" not in result
filters = provider._build_filters()
assert "run_id" not in filters
def test_empty_when_no_params(self, mock_mem0_client: AsyncMock) -> None:
# Validates base query payload generation
provider = Mem0ContextProvider(source_id="mem0", mem0_client=mock_mem0_client)
result = provider._build_search_kwargs("test query", "custom_key", "custom_val")
assert result == {"query": "test query", "filters": {"custom_key": "custom_val"}}
@pytest.mark.asyncio
async def test_before_run_application_only_fallback(self, mock_mem0_client: AsyncMock) -> None:
provider = Mem0ContextProvider(
source_id="mem0", mem0_client=mock_mem0_client, application_id="app_fallback_test"
)
# Mock a valid message list and session container setup
mock_context = MagicMock(spec=SessionContext)
mock_msg = MagicMock()
mock_msg.text = "Retrieve systemic fallback memory traces"
mock_context.input_messages = [mock_msg]
mock_context.response = None
mock_mem0_client.search = AsyncMock(return_value=[{"id": "m1", "memory": "System configuration template"}])
await provider.before_run(
agent=MagicMock(), session=MagicMock(spec=AgentSession), context=mock_context, state={}
)
# Verify that an application-scoped search task executed successfully
assert mock_mem0_client.search.call_count == 1
mock_context.extend_messages.assert_called_once()
assert provider._build_filters() == {}
# -- Context manager tests -----------------------------------------------------
@@ -1997,11 +1997,7 @@ class RawOpenAIChatClient( # type: ignore[misc]
metadata: dict[str, Any] = response.metadata or {}
contents: list[Content] = []
local_shell_tool_name = self._get_local_shell_tool_name(options.get("tools"))
try:
response_outputs = response.output # type: ignore[reportUnknownMemberType]
except AttributeError:
response_outputs = []
for item in response_outputs: # type: ignore[reportUnknownVariableType]
for item in response.output: # type: ignore[reportUnknownMemberType]
match item.type:
# types:
# ParsedResponseOutputMessage[Unknown] |
@@ -788,13 +788,13 @@ class RawOpenAIChatCompletionClient( # type: ignore[misc]
def _get_metadata_from_chat_response(self, response: ChatCompletion) -> dict[str, Any]:
"""Get metadata from a chat response."""
return {
"system_fingerprint": getattr(response, "system_fingerprint", None),
"system_fingerprint": response.system_fingerprint,
}
def _get_metadata_from_streaming_chat_response(self, response: ChatCompletionChunk) -> dict[str, Any]:
"""Get metadata from a streaming chat response."""
return {
"system_fingerprint": getattr(response, "system_fingerprint", None),
"system_fingerprint": response.system_fingerprint,
}
def _get_metadata_from_chat_choice(self, choice: Choice | ChunkChoice) -> dict[str, Any]:
+2 -2
View File
@@ -4,7 +4,7 @@ description = "OpenAI integrations for Microsoft Agent Framework."
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.8.1"
version = "1.8.0"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -23,7 +23,7 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core>=1.8.1,<2",
"agent-framework-core>=1.8.0,<2",
"openai>=1.99.0,<3",
]
+2 -2
View File
@@ -4,7 +4,7 @@ description = "Microsoft Agent Framework for building AI Agents with Python. Thi
authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
readme = "README.md"
requires-python = ">=3.10"
version = "1.8.1"
version = "1.8.0"
license-files = ["LICENSE"]
urls.homepage = "https://aka.ms/agent-framework"
urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -23,7 +23,7 @@ classifiers = [
"Typing :: Typed",
]
dependencies = [
"agent-framework-core[all]==1.8.1",
"agent-framework-core[all]==1.8.0",
]
[dependency-groups]
-1
View File
@@ -7,7 +7,6 @@
"**/demos/**",
"**/_to_delete/**",
"**/05-end-to-end/**",
"**/harness/**",
"**/agent_with_foundry_tracing.py",
"**/azure_responses_client_with_foundry.py"
],
-1
View File
@@ -7,7 +7,6 @@
"**/demos/**",
"**/_to_delete/**",
"**/05-end-to-end/**",
"**/harness/**",
"**/agent_with_foundry_tracing.py",
"**/azure_responses_client_with_foundry.py",
"**/github_copilot/**"
@@ -1,94 +0,0 @@
# Harness Console
A Textual-based terminal UI for running and observing AI agents built with the Agent Framework.
## Quick Start
```python
from console import run_agent_async, build_default_observers
await run_agent_async(
agent=my_agent,
session=my_session,
observers=build_default_observers(),
)
```
See [`harness_research.py`](../harness_research.py) for a complete example.
## Package Structure
```
console/
├── __init__.py # Public API exports
├── harness_console.py # run_agent_async() entry point
├── app.py # HarnessApp (Textual application)
├── app_state.py # HarnessAppState, enums, data types
├── agent_runner.py # HarnessAgentRunner (streaming orchestration)
├── state_driver.py # IUXStateDriver protocol
├── textual_state_driver.py # Textual implementation of IUXStateDriver
├── formatters.py # Tool call formatters
├── observers/ # Lifecycle observers
│ ├── base.py # ConsoleObserver abstract base
│ ├── text_output.py # Streaming text display
│ ├── tool_call_display.py # Tool call formatting
│ ├── tool_approval.py # User approval for tool calls
│ ├── error_display.py # Error messages
│ ├── usage_display.py # Token usage tracking
│ └── reasoning_display.py # Reasoning/thinking blocks
├── components/ # Textual UI widgets
│ ├── scroll_panel.py # Conversation history
│ ├── text_input.py # User text input
│ ├── list_selection.py # Multiple choice selector
│ ├── agent_status.py # Spinner + usage display
│ └── agent_mode_help.py # Mode indicator + help text
└── commands/ # Slash command handlers
├── base.py # CommandHandler abstract base
├── exit_handler.py # /exit
├── mode_handler.py # /mode [plan|execute]
├── todo_handler.py # /todos
└── session_handler.py # /session-export, /session-import
```
## Public API
| Export | Description |
|--------|-------------|
| `run_agent_async` | Main entry point — runs the Textual app with an agent |
| `build_default_observers` | Factory for the standard observer set |
| `build_default_command_handlers` | Factory for slash command handlers |
| `ConsoleObserver` | Base class for custom observers |
| `ToolCallFormatter` | Base class for custom tool formatters |
| `CommandHandler` | Base class for custom slash commands |
## Architecture
The console follows a unidirectional data flow:
```
AgentRunner → Observers → StateDriver → AppState → Textual UI
↑
User Input (app.py)
```
- **AgentRunner** streams responses from the agent and dispatches events to observers.
- **Observers** process events (text chunks, tool calls, errors) and update the state driver.
- **StateDriver** (`IUXStateDriver`) mutates `HarnessAppState` and notifies the UI.
- **Textual App** reads state and syncs widgets on each notification.
### Key Design Choices
| Concern | Approach |
|---------|----------|
| Rendering | Textual widgets + Rich markup (no manual ANSI) |
| State | Single `HarnessAppState` dataclass, mutated by driver |
| Streaming text | Truncate-and-rewrite on RichLog for flicker-free updates |
| Extensibility | Custom observers, formatters, and commands via base classes |
| Follow-up questions | Observer returns `FollowUpQuestion` → UI shows prompt/choices |
## Dependencies
- `textual` — TUI framework
- `rich` — Text formatting
- `agent-framework` — Core agent framework
@@ -1,27 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Harness Console - A Textual-based TUI for AI agent interactions.
This package provides a rich terminal interface for running and observing
AI agents, with streaming output, tool call display, follow-up questions,
and token usage tracking.
"""
from .commands import CommandHandler, build_default_command_handlers
from .formatters import ToolCallFormatter
from .harness_console import run_agent_async
from .observers import (
ConsoleObserver,
build_default_observers,
build_observers_with_planning,
)
__all__ = [
"CommandHandler",
"ConsoleObserver",
"ToolCallFormatter",
"build_default_command_handlers",
"build_default_observers",
"build_observers_with_planning",
"run_agent_async",
]
@@ -1,343 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Agent runner orchestration for the harness console.
This module provides the HarnessAgentRunner class, which orchestrates agent
invocations with observer lifecycle management. It handles:
- User input dispatch
- Agent streaming with observer notifications
- Follow-up action collection
- Streaming state management
"""
from __future__ import annotations
import asyncio
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from agent_framework import Agent, AgentSession
from .app_state import FollowUpAction
from .observers.base import ConsoleObserver
from .state_driver import IUXStateDriver
class HarnessAgentRunner:
"""Orchestrates agent invocations driven by user-input events from the UI.
The component invokes the runner's input handlers (run_turn) directly;
the runner mutates UI state through the supplied IUXStateDriver.
This is a minimal implementation focusing on the core agent loop without
command handling or complex message injection (those can be added later).
"""
def __init__(
self,
agent: Agent,
observers: list[ConsoleObserver],
state_driver: IUXStateDriver,
*,
max_context_window_tokens: int | None = None,
max_output_tokens: int | None = None,
) -> None:
"""Initialize the agent runner.
Args:
agent: The agent to orchestrate.
observers: List of console observers for lifecycle events.
state_driver: The UI state driver for observer updates.
max_context_window_tokens: Optional max context window size for usage display.
max_output_tokens: Optional max output tokens for usage display.
"""
self._agent = agent
self._observers = observers
self._ux = state_driver
self._max_context_window_tokens = max_context_window_tokens
self._max_output_tokens = max_output_tokens
self._input_gate = asyncio.Semaphore(1) # Single turn at a time
async def run_turn(
self,
user_input: str,
session: AgentSession | None = None,
) -> None:
"""Run a single agent turn with the given user input.
Echoes the input, then delegates to the agent loop.
Args:
user_input: The user's input text.
session: Optional agent session for conversation history.
"""
async with self._input_gate:
self._ux.write_user_input_echo(user_input)
from agent_framework import Message
messages = [Message(role="user", contents=[user_input])]
await self._run_agent_loop(messages, session)
async def start_agent_turn(
self,
messages: list,
session: AgentSession | None = None,
) -> None:
"""Resume the agent loop with pre-built messages (from follow-up responses).
Called by the app after the user finishes answering follow-up questions.
If messages is empty, just completes the turn.
Args:
messages: List of Message objects to send to the agent.
session: Optional agent session.
"""
async with self._input_gate:
if not messages:
self._complete_turn()
return
await self._run_agent_loop(messages, session)
async def _run_agent_loop(
self,
messages: list,
session: AgentSession | None,
) -> None:
"""Run the agent loop, re-invoking as needed for follow-up messages.
Loops while there are messages to send. After each stream:
- Collects follow-up actions from observers
- If questions exist → queue them and return (UI will collect answers)
- If only direct messages → loop with those messages
- If nothing → complete the turn
Args:
messages: Initial messages to send.
session: Optional agent session.
"""
next_messages = messages
while next_messages:
# Configure run options
options = self._configure_run_options(session)
# Begin streaming
self._ux.begin_streaming()
self._ux.begin_streaming_output()
self._ux.set_show_spinner(True)
try:
await self._stream_response_messages(next_messages, session, options)
except Exception as ex:
self._ux.append_info_line(
f"❌ Stream error: {ex.__class__.__name__}:\n{ex}",
color="red",
)
# Stop spinner and end streaming output
self._ux.set_show_spinner(False)
# Collect follow-up actions from observers
follow_up_actions = await self._collect_follow_up_actions(session)
# Separate direct messages from questions
has_follow_ups = len(follow_up_actions) > 0
# Write no-text warning if applicable
await self._ux.write_no_text_warning(has_follow_ups)
# Enqueue all follow-up actions
for action in follow_up_actions:
self._ux.enqueue_follow_up_action(action)
# Check if there are pending questions (UI needs user input)
if self._ux.has_pending_questions():
# Pause — the UI will collect answers and call start_agent_turn
return
# No questions — drain any accumulated direct messages and loop
drained = self._ux.take_follow_up_responses()
next_messages = drained if drained else None
self._complete_turn()
def _complete_turn(self) -> None:
"""Complete the current turn (end streaming)."""
self._ux.end_streaming()
def _configure_run_options(
self,
session: AgentSession | None,
) -> dict:
"""Configure run options via observers.
Each observer can modify the options dict to influence agent behavior.
Args:
session: Optional agent session.
Returns:
Options dict for agent.run().
"""
options = {}
for observer in self._observers:
observer.configure_run_options(options, self._agent, session)
return options
async def _stream_response(
self,
user_input: str,
session: AgentSession | None,
options: dict,
) -> None:
"""Stream agent response from a text input and dispatch to observers.
Args:
user_input: The user's input text.
session: Optional agent session.
options: Run options configured by observers.
"""
# Stream response using agent.run(stream=True)
stream = self._agent.run(
user_input,
stream=True,
session=session,
options=options,
)
# Process each update chunk
async for update in stream:
await self._dispatch_update(update, session)
# Extract usage from the final response
self._extract_usage(stream)
async def _stream_response_messages(
self,
messages: list,
session: AgentSession | None,
options: dict,
) -> None:
"""Stream agent response from Message objects and dispatch to observers.
Args:
messages: List of Message objects to send.
session: Optional agent session.
options: Run options configured by observers.
"""
stream = self._agent.run(
messages,
stream=True,
session=session,
options=options,
)
async for update in stream:
await self._dispatch_update(update, session)
self._extract_usage(stream)
def _extract_usage(self, stream) -> None:
"""Extract token usage from a completed stream."""
try:
get_final = getattr(stream, "get_final_response", None)
if not get_final:
return
import inspect
if inspect.iscoroutinefunction(get_final):
return
final_response = get_final()
if final_response is None:
return
usage = getattr(final_response, "usage_details", None)
if not isinstance(usage, dict):
return
input_tokens = usage.get("input_token_count", 0) or 0
output_tokens = usage.get("output_token_count", 0) or 0
if input_tokens or output_tokens:
self._ux.set_usage_text(self._format_usage(input_tokens, output_tokens))
except (AttributeError, TypeError):
pass
async def _dispatch_update(
self,
update, # AgentResponseUpdate
session: AgentSession | None,
) -> None:
"""Dispatch a single update to all observers.
Calls observer lifecycle methods in order:
1. on_response_update (once per update)
2. on_content (for each content item)
3. on_text (if text is present)
Args:
update: The agent response update.
session: Optional agent session.
"""
# on_response_update
for observer in self._observers:
await observer.on_response_update(self._ux, update, self._agent, session)
# on_content for each content item
if hasattr(update, "contents") and update.contents:
for content in update.contents:
for observer in self._observers:
await observer.on_content(self._ux, content, self._agent, session)
# on_text for text chunks
if hasattr(update, "text") and update.text:
for observer in self._observers:
await observer.on_text(self._ux, update.text, self._agent, session)
async def _collect_follow_up_actions(
self,
session: AgentSession | None,
) -> list[FollowUpAction]:
"""Collect follow-up actions from all observers.
Called after streaming completes to gather any follow-up questions
or messages from observers.
Args:
session: Optional agent session.
Returns:
List of follow-up actions from all observers.
"""
actions: list[FollowUpAction] = []
for observer in self._observers:
observer_actions = await observer.on_stream_complete(
self._ux, self._agent, session
)
if observer_actions:
actions.extend(observer_actions)
return actions
def _format_usage(self, input_tokens: int, output_tokens: int) -> str:
"""Format token counts matching C# harness style: 📊 Tokens — input: X | output: Y | total: Z."""
total_tokens = input_tokens + output_tokens
input_budget = None
if self._max_context_window_tokens and self._max_output_tokens:
input_budget = self._max_context_window_tokens - self._max_output_tokens
return (
f"📊 Tokens — input: {self._format_token_count(input_tokens, input_budget)}"
f" | output: {self._format_token_count(output_tokens, self._max_output_tokens)}"
f" | total: {self._format_token_count(total_tokens, self._max_context_window_tokens)}"
)
@staticmethod
def _format_token_count(count: int, budget: int | None) -> str:
"""Format a token count, optionally showing budget percentage."""
if budget and budget > 0:
pct = count / budget * 100
return f"{count:,}/{budget:,} ({pct:.1f}%)"
return f"{count:,}"
@@ -1,541 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Main Textual application for the harness console.
This module provides the HarnessApp - the main Textual application that
composes all UI components and integrates with the agent runner.
"""
from __future__ import annotations
from typing import TYPE_CHECKING
from textual import on, work
from textual.app import App, ComposeResult
from textual.binding import Binding
from textual.containers import Container, Vertical
from textual.css.query import NoMatches
from textual.widgets import Input, Static
from .app_state import (
BottomPanelMode,
HarnessAppState,
OutputEntryType,
)
from .components import (
AgentModeAndHelp,
AgentStatus,
HarnessListSelection,
HarnessScrollPanel,
HarnessTextInput,
PromptRule,
)
from .textual_state_driver import HarnessConsoleUXStateDriver
if TYPE_CHECKING:
from agent_framework import Agent, AgentSession
from .agent_runner import HarnessAgentRunner
from .commands import CommandHandler
from .observers.base import ConsoleObserver
class HarnessApp(App[None]):
"""Main Textual application for the harness console.
Composes the scroll panel (conversation history), status bar (spinner, usage),
mode/help display, and bottom panel (text input, list selection, or streaming
indicator). Routes user input to the agent runner.
"""
CSS = """
Screen {
background: $background;
}
#scroll-panel {
height: 1fr;
padding: 0 1;
background: transparent;
}
#bottom-panel {
height: auto;
}
#text-input-container {
height: 1;
display: block;
}
#list-selection-container {
height: auto;
max-height: 12;
display: none;
}
#streaming-indicator {
height: 1;
display: none;
}
#status-bar {
height: 1;
}
#mode-help {
height: 1;
}
#top-rule {
height: 1;
}
#bottom-rule {
height: 1;
}
#separator-rule {
height: 1;
}
#text-input {
height: 1;
}
.hidden {
display: none;
}
.visible {
display: block;
}
.input-field {
border: none;
padding: 0;
min-height: 1;
height: 1;
background: transparent;
}
.input-field:focus {
border: none;
background: transparent;
}
.prompt-container {
height: 1;
}
.prompt-label {
width: 2;
min-width: 2;
height: 1;
}
"""
BINDINGS = [
Binding("ctrl+c", "quit", "Quit", show=False),
Binding("ctrl+q", "quit", "Quit", show=False),
]
def __init__(
self,
agent: Agent,
observers: list[ConsoleObserver],
session: AgentSession | None = None,
mode_colors: dict[str, str] | None = None,
initial_mode: str | None = None,
placeholder: str = "Type a message and press Enter...",
title: str = "Harness Console",
max_context_window_tokens: int | None = None,
max_output_tokens: int | None = None,
command_handlers: list[CommandHandler] | None = None,
) -> None:
"""Initialize the harness console application.
Args:
agent: The agent to run.
observers: List of console observers.
session: Optional agent session.
mode_colors: Optional mode color mapping.
initial_mode: Initial agent mode.
placeholder: Input placeholder text.
title: Application title.
max_context_window_tokens: Optional max context window tokens for usage display.
max_output_tokens: Optional max output tokens for usage display.
command_handlers: Optional list of command handlers. If None, auto-detected.
"""
super().__init__()
self.title = title
self._agent = agent
self._observers = observers
self._session = session
self._mode_colors = mode_colors
self._initial_mode = initial_mode
self._placeholder = placeholder
self._max_context_window_tokens = max_context_window_tokens
self._max_output_tokens = max_output_tokens
# Build command handlers
if command_handlers is None:
from .commands import build_default_command_handlers
self._command_handlers = build_default_command_handlers(
agent, mode_colors=mode_colors
)
else:
self._command_handlers = command_handlers
# Compute help text from command handlers
help_parts = [
h.get_help_text()
for h in self._command_handlers
if h.get_help_text() is not None
]
help_text = ", ".join(help_parts) if help_parts else None
# State and driver
self._app_state = HarnessAppState(
placeholder=placeholder,
mode_text=initial_mode,
help_text=help_text,
)
self._ux_driver = HarnessConsoleUXStateDriver(
app_state=self._app_state,
on_state_changed=self._on_state_changed,
mode_colors=mode_colors,
)
# Agent runner (created after init)
self._runner: HarnessAgentRunner | None = None
@property
def ux_driver(self) -> HarnessConsoleUXStateDriver:
"""Get the UX state driver."""
return self._ux_driver
@property
def runner(self) -> HarnessAgentRunner | None:
"""Get the agent runner."""
return self._runner
def compose(self) -> ComposeResult:
"""Compose the application layout."""
with Vertical():
# Main scroll panel for conversation history
yield HarnessScrollPanel(id="scroll-panel")
# Blank line separating scroll content from status area
yield Static(" ", id="separator-rule")
# Status bar (spinner + usage)
yield AgentStatus(id="status-bar")
# Top rule (mode-colored)
yield PromptRule(id="top-rule")
# Bottom panel - switches between text input, list selection, streaming
with Container(id="bottom-panel"):
# Text input (default)
with Container(id="text-input-container"):
text_input = HarnessTextInput(id="text-input")
text_input.placeholder = self._placeholder
yield text_input
# List selection (for follow-up questions)
with Container(id="list-selection-container"):
yield HarnessListSelection(id="list-selection")
# Bottom rule (mode-colored)
yield PromptRule(id="bottom-rule")
# Mode and help
yield AgentModeAndHelp(id="mode-help")
def on_mount(self) -> None:
"""Initialize after mount."""
# Create agent runner now that everything is set up
from .agent_runner import HarnessAgentRunner
self._runner = HarnessAgentRunner(
agent=self._agent,
observers=self._observers,
state_driver=self._ux_driver,
max_context_window_tokens=self._max_context_window_tokens,
max_output_tokens=self._max_output_tokens,
)
# Set initial mode
if self._initial_mode:
self._ux_driver.current_mode = self._initial_mode
# Focus the text input
try:
text_input = self.query_one("#text-input", HarnessTextInput)
text_input.focus_input()
except NoMatches:
pass
# Set initial rule colors and mode display
self._sync_mode_help()
# --- Event handlers ---
@on(HarnessTextInput.Submitted)
def on_text_submitted(self, event: HarnessTextInput.Submitted) -> None:
"""Handle text input submission."""
text = event.value.strip()
if not text:
return
if self._app_state.pending_questions:
# Answer the current follow-up question
self._handle_follow_up_answer(text)
elif self._app_state.mode == BottomPanelMode.STREAMING:
# Input during streaming (message injection placeholder)
pass
elif text.startswith("/"):
# Try command handlers
self._try_command_handlers(text)
else:
# Normal user input - run agent turn
self._run_agent_turn(text)
@work(exclusive=True, thread=False)
async def _try_command_handlers(self, text: str) -> None:
"""Try each command handler; fall through to agent if none match."""
session = self._session
if session is None:
# No session — fall through to agent turn
self._run_agent_turn(text)
return
for handler in self._command_handlers:
if await handler.try_handle(text, session, self._ux_driver):
# Command handled — check for shutdown/session swap signals
self._process_command_signals()
return
# No handler matched — treat as normal agent input
self._run_agent_turn(text)
def _process_command_signals(self) -> None:
"""Check and process signals set by command handlers."""
if self._app_state.shutdown_requested:
self.exit()
return
if self._app_state.replaced_session is not None:
self._session = self._app_state.replaced_session # type: ignore[assignment]
self._app_state.replaced_session = None
self._ux_driver.append_info_line("Session replaced.")
self._sync_ui_from_state()
@on(HarnessListSelection.Selected)
def on_list_selected(self, event: HarnessListSelection.Selected) -> None:
"""Handle list selection."""
self._handle_follow_up_answer(event.value)
# --- Agent turn ---
@work(exclusive=True, thread=False)
async def _run_agent_turn(self, text: str) -> None:
"""Run an agent turn in a background worker."""
if self._runner is None:
return
await self._runner.run_turn(text, session=self._session)
# After turn completes, check for follow-up questions
self._sync_ui_from_state()
# --- Follow-up question handling ---
@work(exclusive=True, thread=False)
async def _handle_follow_up_answer(self, answer: str) -> None:
"""Handle a user's answer to a follow-up question."""
if not self._app_state.pending_questions:
return
question = self._app_state.pending_questions[0]
# Call the continuation
result_message = await question.continuation(answer, self._ux_driver)
# Add result to accumulated responses
if result_message is not None:
self._ux_driver.add_follow_up_response(result_message)
# Advance to next question
self._ux_driver.advance_follow_up_question()
# If no more questions, resume the agent with accumulated responses
if not self._app_state.pending_questions:
responses = self._ux_driver.take_follow_up_responses()
if responses and self._runner:
await self._runner.start_agent_turn(responses, session=self._session)
self._sync_ui_from_state()
# --- State synchronization ---
def _on_state_changed(self) -> None:
"""Called by state driver when state changes - schedule UI sync.
Since the agent runner uses @work(thread=False), state changes happen
on the main event loop. We use call_later to batch updates.
"""
self.call_later(self._sync_ui_from_state)
def _sync_ui_from_state(self) -> None:
"""Synchronize UI components with current application state."""
state = self._app_state
# Update scroll panel with new entries
self._sync_scroll_panel()
# Update bottom panel mode
self._sync_bottom_panel(state.mode)
# Hide status bar and mode/help during list selection (matching C#)
is_list_mode = state.mode == BottomPanelMode.LIST_SELECTION
self._sync_chrome_visibility(not is_list_mode)
# Update status bar
self._sync_status_bar()
# Update mode/help display
self._sync_mode_help()
def _sync_scroll_panel(self) -> None:
"""Sync the scroll panel with output entries."""
try:
panel = self.query_one("#scroll-panel", HarnessScrollPanel)
except NoMatches:
return
entries = self._app_state.output_entries
rendered_count = getattr(self, "_rendered_entry_count", 0)
if rendered_count < len(entries):
# There are new entries to render
for entry in entries[rendered_count:]:
if entry.type == OutputEntryType.STREAMING_TEXT:
panel.set_streaming_entry(entry)
else:
# End any active streaming before appending other entry types
panel.end_streaming()
panel.append_entry(entry)
self._rendered_entry_count = len(entries)
elif rendered_count == len(entries) and entries:
# Same count — check if the last entry is a streaming entry that was mutated
last_entry = entries[-1]
if last_entry.type == OutputEntryType.STREAMING_TEXT:
panel.set_streaming_entry(last_entry)
def _sync_bottom_panel(self, mode: BottomPanelMode) -> None:
"""Switch the bottom panel between text input, list, and streaming."""
try:
text_container = self.query_one("#text-input-container")
list_container = self.query_one("#list-selection-container")
except NoMatches:
return
if mode == BottomPanelMode.TEXT_INPUT:
text_container.display = True
list_container.display = False
# Restore focus to text input
try:
text_input = self.query_one("#text-input", HarnessTextInput)
text_input.focus_input()
except NoMatches:
pass
elif mode == BottomPanelMode.LIST_SELECTION:
text_container.display = False
list_container.display = True
self._sync_list_selection()
elif mode == BottomPanelMode.STREAMING:
text_container.display = True
list_container.display = False
def _sync_list_selection(self) -> None:
"""Sync the list selection widget with state."""
try:
list_widget = self.query_one("#list-selection", HarnessListSelection)
except NoMatches:
return
state = self._app_state
list_widget.title = state.list_selection_title or ""
list_widget.options = list(state.list_selection_options)
list_widget.allow_custom_text = state.list_selection_custom_text_placeholder is not None
if state.list_selection_custom_text_placeholder:
try:
custom_input = list_widget.query_one("#custom-input", Input)
custom_input.placeholder = state.list_selection_custom_text_placeholder
except Exception:
pass
# Focus the option list so keyboard navigation works immediately
list_widget.focus_list()
def _sync_status_bar(self) -> None:
"""Sync the status bar with state."""
try:
status = self.query_one("#status-bar", AgentStatus)
except NoMatches:
return
state = self._app_state
status.show_spinner = state.show_spinner
status.usage_text = state.usage_text or ""
def _sync_mode_help(self) -> None:
"""Sync the mode/help display and rule colors with state."""
try:
mode_help = self.query_one("#mode-help", AgentModeAndHelp)
except NoMatches:
return
state = self._app_state
mode_help.mode = state.mode_text or ""
mode_help.mode_color = state.mode_color or "blue"
mode_help.help_text = state.help_text or ""
# Sync rule colors to match mode
color = state.mode_color or "cyan"
try:
top_rule = self.query_one("#top-rule", PromptRule)
top_rule.rule_color = color
except NoMatches:
pass
try:
bottom_rule = self.query_one("#bottom-rule", PromptRule)
bottom_rule.rule_color = color
except NoMatches:
pass
def _sync_chrome_visibility(self, visible: bool) -> None:
"""Show or hide chrome elements (status bar, mode/help).
During list selection mode, these are hidden to give more vertical
space to the scroll panel and list picker.
Args:
visible: Whether chrome elements should be visible.
"""
import contextlib
with contextlib.suppress(NoMatches):
self.query_one("#status-bar", AgentStatus).display = visible
with contextlib.suppress(NoMatches):
self.query_one("#mode-help", AgentModeAndHelp).display = visible
# --- Rendering count tracking ---
_rendered_entry_count: int = 0
@@ -1,260 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Application state and core data types for the harness console.
This module defines enums, dataclasses, follow-up action types, and the
HarnessAppState dataclass which holds all UI state that may change during
application execution. The state driver mutates this state to coordinate
between the agent runner and the Textual UI components.
"""
from __future__ import annotations
from collections.abc import Awaitable, Callable
from dataclasses import dataclass, field
from enum import Enum
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from agent_framework import Message
from .state_driver import IUXStateDriver
# region Enums
class OutputEntryType(Enum):
"""Type of output entry in the console conversation."""
USER_INPUT = "user_input"
"""User input echo (e.g., 'You: hello')."""
STREAMING_TEXT = "streaming_text"
"""In-progress streaming text from the agent (accumulated chunk by chunk)."""
INFO_LINE = "info_line"
"""Informational line (tool calls, errors, usage, approval requests, etc.)."""
STREAM_FOOTER = "stream_footer"
"""Stream footer (e.g., '(no text response from agent)')."""
PENDING_MESSAGE = "pending_message"
"""Pending injected message notification."""
class BottomPanelMode(Enum):
"""Mode of the bottom panel UI."""
TEXT_INPUT = "text_input"
"""Show text input for user messages."""
LIST_SELECTION = "list_selection"
"""Show choice list for user selection."""
STREAMING = "streaming"
"""Show 'streaming...' indicator while agent is generating."""
# endregion
# region Output Entry
@dataclass
class OutputEntry:
"""A single output entry in the console conversation history.
Used internally by the state driver to track conversation output,
including streaming text, tool calls, errors, and user input echoes.
Args:
type: The type of output entry.
text: The text content of the entry.
color: Optional Rich color string (e.g., "cyan", "red", "dim").
"""
type: OutputEntryType
text: str
color: str | None = None
# endregion
# region Follow-Up Actions
class FollowUpAction:
"""Base class for follow-up actions returned by observers.
Follow-up actions describe either a question to ask the user
(via FollowUpQuestion subclasses) or a message to add directly
to the next agent input (FollowUpMessage).
"""
pass
@dataclass
class FollowUpQuestion(FollowUpAction):
"""A question to ask the user with a continuation.
The continuation delegate is invoked with the user's answer and the
UX state driver, and returns an optional Message to add to the next
agent invocation.
Args:
prompt: The question text shown to the user.
continuation: Async function invoked with the user's answer and state driver.
Returns an optional Message to add to the next agent input.
"""
prompt: str
continuation: Callable[[str, IUXStateDriver], Awaitable[Message | None]]
@dataclass
class TextFollowUpQuestion(FollowUpQuestion):
"""A free-form text question.
The user may type any response. This is the base FollowUpQuestion type
with no additional constraints.
"""
pass
@dataclass
class ChoiceFollowUpQuestion(FollowUpQuestion):
"""A multiple choice question.
The user picks from the provided choices, with an optional ability to
enter custom text when allow_custom_text is True.
Args:
prompt: The question text shown to the user.
choices: List of pre-defined choices.
allow_custom_text: If True, the user may type a custom response in
addition to the listed choices.
continuation: Async function invoked with the user's choice/text and
state driver. Returns an optional Message to add to the next agent input.
"""
choices: list[str]
allow_custom_text: bool = False
@dataclass
class FollowUpMessage(FollowUpAction):
"""A message to add directly to the next agent invocation without prompting.
Used when an observer wants to inject a message into the conversation
without user interaction (e.g., automatic tool results, system messages).
Args:
message: The Message to add to the conversation.
"""
message: Message
# endregion
# region Application State
@dataclass
class HarnessAppState:
"""All UI state for the harness console application.
This state is mutated by the UX state driver and read by the Textual
app to update the UI.
"""
# --- Bottom panel mode ---
mode: BottomPanelMode = BottomPanelMode.TEXT_INPUT
"""Which component is shown in the bottom panel."""
# --- Follow-up question queue ---
pending_questions: list[FollowUpQuestion] = field(default_factory=list)
"""Queue of follow-up questions waiting for user answers.
The head ([0]) is the question currently being displayed; subsequent items
are dispatched in order as each is answered.
"""
accumulated_follow_up_responses: list[Message] = field(default_factory=list)
"""Accumulated follow-up response messages collected during the current agent turn.
Both direct FollowUpMessages emitted by observers and continuation results
from answered questions. Consumed by the runner via take_follow_up_responses().
"""
# --- Text input (active in TextInput / Streaming modes) ---
prompt: str = "> "
"""The prompt string for text input mode."""
placeholder: str = ""
"""Placeholder text shown when the input is empty."""
input_text: str = ""
"""The current input text being typed."""
input_enabled: bool = True
"""Whether input is enabled (disabled during streaming without injection)."""
streaming_prompt: str = "(agent is running...)"
"""The prompt to show during streaming when input is disabled."""
# --- List selection (active in ListSelection mode) ---
list_selection_title: str | None = None
"""Title text displayed above the list selection."""
list_selection_options: list[str] = field(default_factory=list)
"""The list selection options."""
list_selection_index: int = 0
"""The highlighted option index in list selection mode."""
list_selection_custom_text_placeholder: str | None = None
"""Placeholder text for the custom text input option in the list."""
list_selection_custom_input_text: str = ""
"""Current text being typed into the list's custom text option."""
# --- Scroll / output area ---
output_entries: list[OutputEntry] = field(default_factory=list)
"""Output entries in the scroll area conversation history."""
queued_items: list[str] = field(default_factory=list)
"""Queued input items to display (pending injected messages)."""
# --- Agent mode + status display ---
mode_color: str | None = None
"""Rich color string for the rule borders and mode label."""
mode_text: str | None = None
"""Current mode name displayed (e.g., 'plan', 'execute')."""
help_text: str | None = None
"""Help text displayed below the bottom rule (available commands)."""
show_spinner: bool = False
"""Whether the agent status spinner is visible."""
usage_text: str | None = None
"""Formatted token usage text to display in the status bar."""
# --- Command handler signals ---
shutdown_requested: bool = False
"""Set to True when /exit is invoked; the app should exit."""
replaced_session: object | None = None
"""When set, the app should swap its session to this AgentSession."""
@@ -1,65 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Command handler package for the harness console.
Provides slash-command handling (e.g., /exit, /mode, /todos, /session-export)
that intercepts user input before it reaches the agent.
"""
from __future__ import annotations
from typing import TYPE_CHECKING
from .base import CommandHandler
from .exit_handler import ExitCommandHandler
from .mode_handler import ModeCommandHandler
from .session_handler import SessionCommandHandler
from .todo_handler import TodoCommandHandler
if TYPE_CHECKING:
from agent_framework import Agent
__all__ = [
"CommandHandler",
"ExitCommandHandler",
"ModeCommandHandler",
"SessionCommandHandler",
"TodoCommandHandler",
"build_default_command_handlers",
]
def build_default_command_handlers(
agent: Agent,
*,
mode_colors: dict[str, str] | None = None,
) -> list[CommandHandler]:
"""Build the default set of command handlers by inspecting the agent.
Auto-detects TodoProvider and AgentModeProvider from the agent's
context_providers list.
Args:
agent: The agent to inspect for providers.
mode_colors: Optional mapping of mode names to Rich color strings.
Returns:
List of command handlers in evaluation order.
"""
from agent_framework import AgentModeProvider, TodoProvider
todo_provider: TodoProvider | None = None
mode_provider: AgentModeProvider | None = None
for provider in getattr(agent, "context_providers", []):
if isinstance(provider, TodoProvider) and todo_provider is None:
todo_provider = provider
elif isinstance(provider, AgentModeProvider) and mode_provider is None:
mode_provider = provider
return [
ExitCommandHandler(),
TodoCommandHandler(todo_provider),
ModeCommandHandler(mode_provider, mode_colors),
SessionCommandHandler(),
]
@@ -1,58 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Abstract base class for console command handlers.
Command handlers intercept user input starting with '/' and execute
local commands before input reaches the agent. They are checked in order;
the first handler that accepts the input prevents further handlers from
being checked.
"""
from __future__ import annotations
from abc import ABC, abstractmethod
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from agent_framework import AgentSession
from ..state_driver import IUXStateDriver
class CommandHandler(ABC):
"""Base class for console command handlers.
Subclasses implement get_help_text() for the mode bar and
try_handle() to intercept matching commands.
"""
@abstractmethod
def get_help_text(self) -> str | None:
"""Get the help text for this command.
Displayed in the mode-and-help bar. Return None if the
command is not currently available.
Returns:
Help text like '/todos (show todo list)', or None.
"""
...
@abstractmethod
async def try_handle(
self,
user_input: str,
session: AgentSession,
ux: IUXStateDriver,
) -> bool:
"""Attempt to handle the given user input.
Args:
user_input: The raw user input string.
session: The current agent session.
ux: The UX state driver for rendering output.
Returns:
True if this handler handled the input; False otherwise.
"""
...
@@ -1,35 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Exit command handler — /exit to quit the console."""
from __future__ import annotations
from typing import TYPE_CHECKING
from .base import CommandHandler
if TYPE_CHECKING:
from agent_framework import AgentSession
from ..state_driver import IUXStateDriver
class ExitCommandHandler(CommandHandler):
"""Handle the /exit command to shut down the console application."""
def get_help_text(self) -> str | None:
"""Return help text for the exit command."""
return "/exit (quit)"
async def try_handle(
self,
user_input: str,
session: AgentSession,
ux: IUXStateDriver,
) -> bool:
"""Handle /exit by requesting shutdown."""
if user_input.strip().lower() != "/exit":
return False
ux.request_shutdown()
return True
@@ -1,81 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Mode command handler — /mode to show or switch agent mode."""
from __future__ import annotations
from typing import TYPE_CHECKING
from .base import CommandHandler
if TYPE_CHECKING:
from agent_framework import AgentModeProvider, AgentSession
from ..state_driver import IUXStateDriver
class ModeCommandHandler(CommandHandler):
"""Handle the /mode command to display or switch the current agent mode."""
def __init__(
self,
mode_provider: AgentModeProvider | None,
mode_colors: dict[str, str] | None = None,
) -> None:
"""Initialize with mode provider and color mapping.
Args:
mode_provider: The mode provider, or None if not available.
mode_colors: Optional mapping of mode names to Rich color strings.
"""
self._mode_provider = mode_provider
self._mode_colors = mode_colors or {}
def get_help_text(self) -> str | None:
"""Return help text, or None if mode provider is unavailable."""
if self._mode_provider is None:
return None
return "/mode [plan|execute] (show or switch mode)"
async def try_handle(
self,
user_input: str,
session: AgentSession,
ux: IUXStateDriver,
) -> bool:
"""Handle /mode [name] command."""
stripped = user_input.strip()
lower = stripped.lower()
if not (lower == "/mode" or lower.startswith("/mode ")):
return False
if self._mode_provider is None:
ux.append_info_line("AgentModeProvider is not available.")
return True
parts = stripped.split(None, 1)
if len(parts) < 2:
# Show current mode
from agent_framework import get_agent_mode
current = get_agent_mode(session)
ux.append_info_line(f"Current mode: {current}")
return True
# Switch mode
new_mode = parts[1].strip()
try:
from agent_framework import set_agent_mode
normalized = set_agent_mode(session, new_mode)
color = self._mode_colors.get(normalized)
ux.set_mode(normalized, color)
ux.append_info_line(
f"Switched to {normalized} mode.",
color=color,
)
except ValueError as ex:
ux.append_info_line(str(ex), color="red")
return True
@@ -1,107 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Session command handler — /session-export and /session-import."""
from __future__ import annotations
import json
from typing import TYPE_CHECKING
from .base import CommandHandler
if TYPE_CHECKING:
from agent_framework import AgentSession
from ..state_driver import IUXStateDriver
class SessionCommandHandler(CommandHandler):
"""Handle /session-export and /session-import commands."""
def get_help_text(self) -> str | None:
"""Return help text for session commands."""
return "/session-export <file> | /session-import <file>"
async def try_handle(
self,
user_input: str,
session: AgentSession,
ux: IUXStateDriver,
) -> bool:
"""Handle session export/import commands."""
stripped = user_input.strip()
command = stripped.split(None, 1)[0].lower() if stripped else ""
if command == "/session-export":
await self._handle_export(stripped, session, ux)
return True
if command == "/session-import":
await self._handle_import(stripped, ux)
return True
return False
async def _handle_export(
self,
user_input: str,
session: AgentSession,
ux: IUXStateDriver,
) -> None:
"""Export the current session to a JSON file."""
parts = user_input.split(None, 1)
if len(parts) < 2:
ux.append_info_line("Usage: /session-export <filename>")
return
filename = parts[1].strip()
try:
serialized = session.to_dict()
json_str = json.dumps(serialized, indent=2)
self._write_file(filename, json_str)
ux.append_info_line(f"Session exported to {filename}")
except Exception as ex:
ux.append_info_line(
f"Failed to export session to {filename}: {ex}",
color="red",
)
async def _handle_import(
self,
user_input: str,
ux: IUXStateDriver,
) -> None:
"""Import a session from a JSON file."""
parts = user_input.split(None, 1)
if len(parts) < 2:
ux.append_info_line("Usage: /session-import <filename>")
return
filename = parts[1].strip()
try:
from agent_framework import AgentSession
json_str = self._read_file(filename)
data = json.loads(json_str)
new_session = AgentSession.from_dict(data)
ux.replace_session(new_session)
ux.append_info_line(f"Session imported from {filename}")
except FileNotFoundError:
ux.append_info_line(f"File not found: {filename}", color="red")
except Exception as ex:
ux.append_info_line(
f"Failed to import session from {filename}: {ex}",
color="red",
)
@staticmethod
def _write_file(filename: str, content: str) -> None:
"""Write content to a file (sync helper to satisfy ASYNC230)."""
with open(filename, "w", encoding="utf-8") as f: # noqa: ASYNC230
f.write(content)
@staticmethod
def _read_file(filename: str) -> str:
"""Read content from a file (sync helper to satisfy ASYNC230)."""
with open(filename, encoding="utf-8") as f: # noqa: ASYNC230
return f.read()
@@ -1,66 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Todo command handler — /todos to display the todo list."""
from __future__ import annotations
from typing import TYPE_CHECKING
from .base import CommandHandler
if TYPE_CHECKING:
from agent_framework import AgentSession, TodoProvider
from ..state_driver import IUXStateDriver
class TodoCommandHandler(CommandHandler):
"""Handle the /todos command to display the current todo list."""
def __init__(self, todo_provider: TodoProvider | None) -> None:
"""Initialize with the todo provider.
Args:
todo_provider: The todo provider, or None if not available.
"""
self._todo_provider = todo_provider
def get_help_text(self) -> str | None:
"""Return help text, or None if todo provider is unavailable."""
if self._todo_provider is None:
return None
return "/todos (show todo list)"
async def try_handle(
self,
user_input: str,
session: AgentSession,
ux: IUXStateDriver,
) -> bool:
"""Handle /todos by displaying the todo list."""
if user_input.strip().lower() != "/todos":
return False
if self._todo_provider is None:
ux.append_info_line("TodoProvider is not available.")
return True
todos = await self._todo_provider.store.load_items(
session, source_id=self._todo_provider.source_id
)
if not todos:
ux.append_info_line("No todos yet.")
return True
ux.append_info_line("── Todo List ──")
for item in todos:
status = "âś“" if item.is_complete else "â—‹"
color = "dim" if item.is_complete else None
description = f" — {item.description}" if item.description else ""
ux.append_info_line(
f"[{status}] #{item.id} {item.title}{description}",
color=color,
)
return True
@@ -1,23 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""UI components for the harness console.
This module provides Textual widgets for building the harness console UI,
including status displays, input fields, choice selectors, and scrolling panels.
"""
from .agent_status import AgentStatus
from .list_selection import HarnessListSelection
from .mode_help import AgentModeAndHelp
from .prompt_rule import PromptRule
from .scroll_panel import HarnessScrollPanel
from .text_input import HarnessTextInput
__all__ = [
"AgentStatus",
"AgentModeAndHelp",
"HarnessListSelection",
"PromptRule",
"HarnessScrollPanel",
"HarnessTextInput",
]
@@ -1,66 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Agent status widget with spinner animation and usage statistics."""
from __future__ import annotations
from textual.reactive import reactive
from textual.widgets import Static
class AgentStatus(Static):
"""Agent status bar with animated spinner and token usage display.
Displays an animated braille pattern spinner when the agent is active,
along with token usage statistics. The component automatically updates
the spinner animation at ~10fps for smooth visual feedback.
Attributes:
show_spinner: Whether to display the animated spinner.
usage_text: Token usage text to display (e.g., "1.2K in / 856 out").
"""
# Braille pattern spinner frames for smooth animation
SPINNER_FRAMES = ["â ‹", "â ™", "â ą", "â ¸", "â Ľ", "â ´", "â ¦", "â §", "â ‡", "â Ź"]
show_spinner: reactive[bool] = reactive(False)
usage_text: reactive[str] = reactive("")
def __init__(self, **kwargs) -> None:
"""Initialize the agent status widget."""
super().__init__(**kwargs)
self._spinner_index = 0
def on_mount(self) -> None:
"""Start the spinner animation timer when the widget is mounted."""
# Update spinner at ~10fps (every 0.1 seconds)
self.set_interval(0.1, self._advance_spinner)
def _advance_spinner(self) -> None:
"""Advance the spinner to the next frame."""
if self.show_spinner:
self._spinner_index = (self._spinner_index + 1) % len(self.SPINNER_FRAMES)
self.refresh()
def render(self) -> str:
"""Render the status bar with spinner and usage text.
Returns:
Formatted string with Rich markup for spinner and usage display.
"""
if not self.show_spinner and not self.usage_text:
return ""
parts = []
if self.show_spinner:
frame = self.SPINNER_FRAMES[self._spinner_index]
parts.append(f"[cyan]{frame}[/cyan]")
else:
# Keep consistent spacing when spinner is off
parts.append(" ")
if self.usage_text:
parts.append(f"[dim]{self.usage_text}[/dim]")
return " ".join(parts)
@@ -1,269 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""List selection widget with optional custom text input."""
from __future__ import annotations
from textual import on
from textual.app import ComposeResult
from textual.binding import Binding
from textual.containers import Container
from textual.css.query import NoMatches
from textual.events import Key
from textual.message import Message
from textual.reactive import reactive
from textual.widget import Widget
from textual.widgets import Input, Label, OptionList
from textual.widgets.option_list import Option
class HarnessListSelection(Widget):
"""List selection widget with numbered choices and optional custom text input.
Displays a title, a list of numbered choices that can be selected via
keyboard navigation or number keys (1-9), and an optional custom text
input field at the bottom.
All child nodes (title label, option list, custom input) are always
present in the DOM; visibility is toggled via reactive watchers.
Navigation:
- Down arrow on last list item moves focus to the custom text input
- Up arrow on the custom text input moves focus back to the option list
- When custom input has focus, the option list highlight is cleared
Attributes:
title: The title text displayed above the options.
options: List of option strings to display.
allow_custom_text: Whether to show a custom text input field.
"""
DEFAULT_CSS = """
HarnessListSelection {
height: auto;
max-height: 12;
}
HarnessListSelection .list-selection-container {
height: auto;
}
HarnessListSelection #selection-title {
height: auto;
color: $text;
text-style: bold;
padding: 0 0 0 0;
}
HarnessListSelection #option-list {
height: auto;
max-height: 8;
border: none;
padding: 0;
}
HarnessListSelection #custom-input {
height: auto;
min-height: 1;
margin-top: 0;
border: tall transparent;
}
HarnessListSelection #custom-input:focus {
border: tall $accent;
}
"""
BINDINGS = [
Binding("1", "select_option(0)", "Select option 1", show=False),
Binding("2", "select_option(1)", "Select option 2", show=False),
Binding("3", "select_option(2)", "Select option 3", show=False),
Binding("4", "select_option(3)", "Select option 4", show=False),
Binding("5", "select_option(4)", "Select option 5", show=False),
Binding("6", "select_option(5)", "Select option 6", show=False),
Binding("7", "select_option(6)", "Select option 7", show=False),
Binding("8", "select_option(7)", "Select option 8", show=False),
Binding("9", "select_option(8)", "Select option 9", show=False),
]
title: reactive[str] = reactive("")
options: reactive[list[str]] = reactive(list, always_update=True)
allow_custom_text: reactive[bool] = reactive(False)
class Selected(Message):
"""Message sent when an option is selected.
Attributes:
value: The selected option text or custom text.
"""
def __init__(self, value: str) -> None:
"""Initialize the Selected message.
Args:
value: The selected option text or custom text.
"""
self.value = value
super().__init__()
def compose(self) -> ComposeResult:
"""Compose the widget — all nodes are always present.
Yields:
Title label (hidden if empty), option list, custom input (hidden by default).
"""
with Container(classes="list-selection-container"):
yield Label("", id="selection-title")
yield OptionList(id="option-list")
yield Input(
placeholder="Or type a custom response...",
id="custom-input",
)
def on_mount(self) -> None:
"""Configure initial visibility after mount."""
title_label = self.query_one("#selection-title", Label)
title_label.display = bool(self.title)
custom_input = self.query_one("#custom-input", Input)
custom_input.display = self.allow_custom_text
self._update_options()
def on_key(self, event: Key) -> None:
"""Handle key navigation between option list and custom input.
Args:
event: The key event.
"""
if not self.allow_custom_text:
return
option_list = self.query_one("#option-list", OptionList)
custom_input = self.query_one("#custom-input", Input)
# Down arrow on last item → move to custom input
if event.key == "down" and option_list.has_focus:
last_index = option_list.option_count - 1
if last_index >= 0 and option_list.highlighted == last_index:
option_list.highlighted = None # type: ignore[assignment]
custom_input.focus()
event.prevent_default()
event.stop()
# Up arrow on custom input → move back to option list (last item)
elif event.key == "up" and custom_input.has_focus:
last_index = option_list.option_count - 1
if last_index >= 0:
option_list.highlighted = last_index
option_list.focus()
event.prevent_default()
event.stop()
@on(Input.Changed, "#custom-input")
def on_custom_input_focused_or_changed(self, event: Input.Changed) -> None:
"""Clear option list highlight when user is typing in custom input.
Args:
event: The input changed event.
"""
option_list = self.query_one("#option-list", OptionList)
option_list.highlighted = None # type: ignore[assignment]
def watch_title(self, new_title: str) -> None:
"""Update the title label when the title changes.
Args:
new_title: The new title text.
"""
try:
label = self.query_one("#selection-title", Label)
label.update(new_title)
label.display = bool(new_title)
except NoMatches:
pass
def watch_options(self, new_options: list[str]) -> None:
"""Update the option list when options change.
Args:
new_options: The new list of options.
"""
import contextlib
with contextlib.suppress(NoMatches):
self._update_options()
def watch_allow_custom_text(self, allow: bool) -> None:
"""Show/hide the custom input field.
Args:
allow: Whether to show the custom text input.
"""
try:
custom_input = self.query_one("#custom-input", Input)
custom_input.display = allow
except NoMatches:
pass
def _update_options(self) -> None:
"""Update the OptionList with numbered options."""
try:
option_list = self.query_one("#option-list", OptionList)
option_list.clear_options()
for i, option_text in enumerate(self.options):
display_text = f"{i + 1}. {option_text}" if i < 9 else f" {option_text}"
option_list.add_option(Option(display_text, id=str(i)))
except NoMatches:
pass
@on(OptionList.OptionSelected)
def on_option_selected(self, event: OptionList.OptionSelected) -> None:
"""Handle option selection from the list.
Args:
event: The OptionList.OptionSelected event.
"""
option_index = int(event.option.id or "0")
if 0 <= option_index < len(self.options):
selected_value = self.options[option_index]
self.post_message(self.Selected(selected_value))
@on(Input.Submitted)
def on_input_submitted(self, event: Input.Submitted) -> None:
"""Handle custom text input submission.
Args:
event: The Input.Submitted event.
"""
if self.allow_custom_text and event.value:
self.post_message(self.Selected(event.value))
event.input.clear()
def action_select_option(self, index: int) -> None:
"""Select an option by index (0-based).
Args:
index: The option index to select.
"""
if 0 <= index < len(self.options):
selected_value = self.options[index]
self.post_message(self.Selected(selected_value))
def focus_list(self) -> None:
"""Focus the option list."""
try:
option_list = self.query_one("#option-list", OptionList)
option_list.focus()
except NoMatches:
pass
def focus_custom_input(self) -> None:
"""Focus the custom text input field."""
if self.allow_custom_text:
try:
custom_input = self.query_one("#custom-input", Input)
custom_input.focus()
except NoMatches:
pass
@@ -1,48 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Agent mode and help text display widget."""
from __future__ import annotations
from rich.text import Text
from textual.reactive import reactive
from textual.widgets import Static
class AgentModeAndHelp(Static):
"""Widget displaying the current agent mode and help text.
Shows the current agent mode (e.g., "plan", "execute") in a colored label,
followed by available commands and help text in a dimmed style. Used in
the fixed bottom area of the console.
Attributes:
mode: Current mode name (e.g., "plan", "execute"), or None if no mode.
mode_color: Rich color string for the mode label (e.g., "yellow", "green").
help_text: Help text to display (e.g., "/exit to quit, /mode to switch").
"""
mode: reactive[str | None] = reactive(None)
mode_color: reactive[str] = reactive("yellow")
help_text: reactive[str] = reactive("")
def render(self) -> Text:
"""Render the mode indicator and help text.
Returns:
Rich Text object with styled mode and help display.
"""
result = Text()
if self.mode:
result.append(f"[{self.mode}]", style=self.mode_color)
if self.help_text:
if self.mode:
result.append(" ")
result.append(self.help_text, style="dim")
if not result.plain:
result.append(" ")
return result
@@ -1,31 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Mode-colored horizontal rule."""
from __future__ import annotations
from textual.reactive import reactive
from textual.widgets import Static
class PromptRule(Static):
"""A full-width horizontal rule colored by the current agent mode.
Renders a line of '─' characters across the terminal width,
colored to match the current mode (e.g., cyan for plan, green for execute).
Attributes:
rule_color: Rich color string for the rule (e.g., "cyan", "green").
"""
rule_color: reactive[str] = reactive("cyan")
def render(self) -> str:
"""Render the horizontal rule.
Returns:
Formatted string with Rich markup.
"""
color = self.rule_color
width = self.size.width or 80
return f"[{color}]{'─' * width}[/{color}]"
@@ -1,127 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Scrolling panel for conversation history display."""
from __future__ import annotations
from typing import TYPE_CHECKING
from textual.widgets import RichLog
if TYPE_CHECKING:
from ..app_state import OutputEntry
class HarnessScrollPanel(RichLog):
"""Scrolling panel for displaying conversation history.
Uses Textual's RichLog widget for efficient append-only rendering with
Rich text formatting support. Automatically scrolls to the bottom when
new entries are added.
For streaming text, the panel uses a truncate-and-rewrite strategy: it
tracks where streaming began in the RichLog lines list, and on each update
truncates back to that point and rewrites the full accumulated text as a
single write. This ensures consistent rendering without line-break artifacts
between streamed chunks.
"""
def __init__(self, **kwargs) -> None:
"""Initialize the scroll panel.
Args:
**kwargs: Additional arguments passed to RichLog.
"""
super().__init__(
**kwargs,
auto_scroll=True, # Automatically scroll to bottom
wrap=True, # Wrap long lines instead of horizontal scroll
markup=True, # Enable Rich markup
highlight=True, # Enable syntax highlighting
)
self._entries: list[OutputEntry] = []
self._is_streaming = False
self._streaming_line_start: int = 0
def append_entry(self, entry: OutputEntry) -> None:
"""Append a new output entry to the conversation history.
Args:
entry: The output entry to append.
"""
self._entries.append(entry)
text = self._format_entry(entry)
self.write(text)
def set_streaming_entry(self, entry: OutputEntry) -> None:
"""Set or update the current streaming entry.
On each update, truncates the RichLog back to where streaming
started, then rewrites the full streaming text as a single block.
This ensures no spurious line breaks between chunks while avoiding
a full rewrite of all entries.
Args:
entry: The streaming entry (will be mutated externally).
"""
if not self._is_streaming:
# First streaming chunk — record where streaming lines begin
self._is_streaming = True
self._entries.append(entry)
self._streaming_line_start = len(self.lines)
# Truncate lines back to where streaming started
if len(self.lines) > self._streaming_line_start:
del self.lines[self._streaming_line_start:]
from textual.geometry import Size
self.virtual_size = Size(self._widest_line_width, len(self.lines))
# Write full streaming text as a single renderable
formatted = self._format_text(entry.text, entry.color)
self.write(formatted)
def end_streaming(self) -> None:
"""End the current streaming mode."""
if self._is_streaming:
self._is_streaming = False
self._streaming_line_start = 0
def _rewrite_all(self) -> None:
"""Clear and rewrite all entries from scratch."""
self.clear()
for entry in self._entries:
self.write(self._format_entry(entry))
def _format_entry(self, entry: OutputEntry) -> str:
"""Format an output entry with Rich markup.
Args:
entry: The entry to format.
Returns:
Formatted string with Rich markup for color and styling.
"""
return self._format_text(entry.text, entry.color)
@staticmethod
def _format_text(text: str, color: str | None) -> str:
"""Format text with optional Rich color markup.
Args:
text: The text to format.
color: Optional Rich color name.
Returns:
Formatted string.
"""
if color:
return f"[{color}]{text}[/{color}]"
return text
def clear_history(self) -> None:
"""Clear all conversation history from the panel."""
self._entries.clear()
self._is_streaming = False
self._streaming_line_start = 0
self.clear()
@@ -1,102 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Text input widget with inline prompt for the harness console."""
from __future__ import annotations
from textual import on
from textual.app import ComposeResult
from textual.containers import Horizontal
from textual.message import Message
from textual.reactive import reactive
from textual.widget import Widget
from textual.widgets import Input, Label
class HarnessTextInput(Widget):
"""Text input widget with a prompt label on the left.
Displays a prompt (e.g., "> ") followed by a borderless input field.
Sits between the two mode-colored horizontal rules.
Attributes:
prompt: The prompt text displayed on the left (e.g., "> ").
placeholder: Placeholder text shown when the input is empty.
"""
prompt: reactive[str] = reactive("> ")
placeholder: reactive[str] = reactive("")
class Submitted(Message):
"""Message sent when the input is submitted.
Attributes:
value: The submitted text value.
"""
def __init__(self, value: str) -> None:
"""Initialize the Submitted message.
Args:
value: The submitted text value.
"""
self.value = value
super().__init__()
def compose(self) -> ComposeResult:
"""Compose the prompt label and input field.
Yields:
A horizontal container with the prompt and input field.
"""
with Horizontal(classes="prompt-container"):
yield Label(self.prompt, classes="prompt-label", id="prompt-label")
yield Input(placeholder=self.placeholder, classes="input-field", id="input-field")
def watch_prompt(self, new_prompt: str) -> None:
"""Update the prompt label when the prompt attribute changes.
Args:
new_prompt: The new prompt text.
"""
try:
label = self.query_one("#prompt-label", Label)
label.update(new_prompt)
except Exception:
pass
def watch_placeholder(self, new_placeholder: str) -> None:
"""Update the input placeholder when the placeholder attribute changes.
Args:
new_placeholder: The new placeholder text.
"""
try:
input_field = self.query_one("#input-field", Input)
input_field.placeholder = new_placeholder
except Exception:
# Input doesn't exist yet (before compose), ignore
pass
@on(Input.Submitted)
def on_input_submitted(self, event: Input.Submitted) -> None:
"""Handle input submission.
Clears the input field and posts a Submitted message with the value.
Args:
event: The Input.Submitted event.
"""
value = event.value
event.input.clear()
self.post_message(self.Submitted(value))
def focus_input(self) -> None:
"""Focus the input field."""
input_field = self.query_one(".input-field", Input)
input_field.focus()
def clear_input(self) -> None:
"""Clear the input field."""
input_field = self.query_one(".input-field", Input)
input_field.clear()
@@ -1,503 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Tool call formatters for displaying function calls in the harness console.
This module provides formatters that convert raw function call content into
human-readable display strings. Each formatter handles specific tool patterns
(e.g., web_search, todos_*, etc.) and the FallbackToolFormatter provides
generic formatting for any unmatched tools.
Usage:
from harness.console.formatters import build_default_formatters, format_tool_call
from agent_framework import Content
call = Content.from_function_call(
call_id="call_1",
name="web_search",
arguments={"query": "Python async"}
)
formatters = build_default_formatters()
result = format_tool_call(formatters, call) # "web_search (Python async)"
"""
from __future__ import annotations
import contextlib
import json
from abc import ABC, abstractmethod
from typing import Any
from agent_framework import Content
# region Helper Functions
def get_argument_value(call: Content, param_name: str) -> Any:
"""Extract an argument value from a function call.
Handles both dict and JSON string arguments.
Args:
call: The function call content.
param_name: The parameter name to extract.
Returns:
The argument value, or None if not found.
"""
if call.arguments is None:
return None
if isinstance(call.arguments, str):
# arguments is a JSON string, parse it
try:
args_dict = json.loads(call.arguments)
except (json.JSONDecodeError, TypeError):
return None
if not isinstance(args_dict, dict):
return None
elif isinstance(call.arguments, dict):
args_dict = call.arguments
else:
return None
return args_dict.get(param_name)
def as_int_list(value: Any) -> list[int] | None:
"""Convert a value to a list of integers, or None if not possible.
Args:
value: The value to convert (should be a list).
Returns:
A list of integers, or None if conversion fails.
"""
if not isinstance(value, list):
return None
result: list[int] = []
for item in value:
if isinstance(item, int):
result.append(item)
else:
with contextlib.suppress(ValueError, TypeError):
result.append(int(item))
return result if result else None
def as_dict_list(value: Any) -> list[dict[str, Any]] | None:
"""Convert a value to a list of dicts, or None if not possible.
Args:
value: The value to convert (should be a list).
Returns:
A list of dicts, or None if value is not a list of dicts.
"""
if not isinstance(value, list):
return None
result: list[dict[str, Any]] = []
for item in value:
if isinstance(item, dict):
result.append(item)
return result if result else None
def truncate(text: str, max_length: int) -> str:
"""Truncate a string to the specified maximum length, appending an ellipsis if truncated.
Args:
text: The text to truncate.
max_length: The maximum length.
Returns:
The truncated string.
"""
return text if len(text) <= max_length else text[:max_length] + "…"
# endregion
# region Base Class
class ToolCallFormatter(ABC):
"""Base class for tool call formatters that produce human-readable display strings
for function call content items shown in the console.
"""
@abstractmethod
def can_format(self, call: Content) -> bool:
"""Return True if this formatter can handle the given function call.
Args:
call: The function call content to check.
Returns:
True if this formatter should be used; otherwise False.
"""
...
@abstractmethod
def format_detail(self, call: Content) -> str | None:
"""Return the detail portion of the formatted output for the given tool call,
or None if only the tool name should be displayed.
Args:
call: The function call content to format.
Returns:
A detail string to append after the tool name, or None.
"""
...
# endregion
# region Concrete Formatters
class FallbackToolFormatter(ToolCallFormatter):
"""Catch-all formatter that handles any tool not matched by a more specific formatter.
Displays a generic summary of the tool's arguments. This formatter should always be
placed last in the formatter list.
"""
def can_format(self, call: Content) -> bool:
"""Always returns True - this formatter matches everything."""
return True
def format_detail(self, call: Content) -> str | None:
"""Format arguments as generic (key: value, ...) pairs."""
if call.arguments is None:
return None
# Parse arguments
if isinstance(call.arguments, str):
try:
args_dict = json.loads(call.arguments)
except (json.JSONDecodeError, TypeError):
return None
elif isinstance(call.arguments, dict):
args_dict = call.arguments
else:
return None
if not args_dict:
return None
# Build argument list
parts: list[str] = []
for key, value in args_dict.items():
if value is None:
continue
# Convert value to string
if isinstance(value, bool):
str_value = "true" if value else "false"
elif isinstance(value, (int, float)):
str_value = str(value)
elif isinstance(value, str):
str_value = value
else:
# Complex types - skip for now
continue
parts.append(f"{key}: {truncate(str_value, 40)}")
return f"({', '.join(parts)})" if parts else None
class WebSearchToolFormatter(ToolCallFormatter):
"""Formats web_search tool calls, showing the search query."""
def can_format(self, call: Content) -> bool:
"""Match web_search tool calls."""
return call.name == "web_search"
def format_detail(self, call: Content) -> str | None:
"""Extract and format the query parameter."""
value = get_argument_value(call, "query")
return f"({value})" if value else None
class TodoToolFormatter(ToolCallFormatter):
"""Formats todos_* tool calls with tree-view output for added items
and structured output for complete/remove operations.
"""
def can_format(self, call: Content) -> bool:
"""Match todos_* tool calls."""
return call.name is not None and call.name.startswith("todos_")
def format_detail(self, call: Content) -> str | None:
"""Format based on the specific todos operation."""
if call.name == "todos_add":
return self._format_add_todos(call)
if call.name == "todos_complete":
return self._format_complete_todos(call)
if call.name == "todos_remove":
return self._format_id_list(call, "ids", "Remove")
return None
def _format_add_todos(self, call: Content) -> str | None:
"""Format todos_add with tree view of titles."""
todos = as_dict_list(get_argument_value(call, "todos"))
if not todos:
return None
titles: list[str] = []
for todo in todos:
title = todo.get("title")
if title and isinstance(title, str):
titles.append(title)
if not titles:
return None
# Build tree view
count = len(titles)
plural = "s" if count != 1 else ""
lines = [f"({count} item{plural})"]
for i, title in enumerate(titles):
connector = "├─" if i < count - 1 else "└─"
lines.append(f"\n {connector} {title}")
return "".join(lines)
def _format_complete_todos(self, call: Content) -> str | None:
"""Format todos_complete with tree view of IDs and reasons."""
items = as_dict_list(get_argument_value(call, "items"))
if not items:
return None
entries: list[tuple[int, str | None]] = []
for item in items:
todo_id = item.get("id")
if not isinstance(todo_id, int):
continue
reason = item.get("reason")
reason_str = str(reason) if reason is not None and not isinstance(reason, str) else reason
entries.append((todo_id, reason_str))
if not entries:
return None
# Build tree view
lines: list[str] = []
for i, (todo_id, reason) in enumerate(entries):
connector = "├─" if i < len(entries) - 1 else "└─"
line = f"\n {connector} Complete #{todo_id}"
if reason:
line += f" — {truncate(reason, 80)}"
lines.append(line)
return "".join(lines)
def _format_id_list(self, call: Content, param_name: str, verb: str) -> str | None:
"""Format a list of IDs with a verb (e.g., Remove #1, Remove #2)."""
ids = as_int_list(get_argument_value(call, param_name))
if not ids:
return None
lines: list[str] = []
for i, todo_id in enumerate(ids):
connector = "├─" if i < len(ids) - 1 else "└─"
lines.append(f"\n {connector} {verb} #{todo_id}")
return "".join(lines)
class ModeToolFormatter(ToolCallFormatter):
"""Formats AgentMode_* tool calls, showing the target mode for Set operations."""
def can_format(self, call: Content) -> bool:
"""Match AgentMode_* tool calls."""
return call.name is not None and call.name.startswith("AgentMode_")
def format_detail(self, call: Content) -> str | None:
"""Format based on the specific AgentMode operation."""
if call.name == "AgentMode_Set":
value = get_argument_value(call, "mode")
return f"({value})" if value else None
return None
class BackgroundAgentToolFormatter(ToolCallFormatter):
"""Formats BackgroundAgents_* tool calls with human-readable details
for task start, continue, wait, and result retrieval operations.
"""
def can_format(self, call: Content) -> bool:
"""Match BackgroundAgents_* tool calls."""
return call.name is not None and call.name.startswith("BackgroundAgents_")
def format_detail(self, call: Content) -> str | None:
"""Format based on the specific BackgroundAgents operation."""
if call.name == "BackgroundAgents_StartTask":
return self._format_start_background_task(call)
if call.name == "BackgroundAgents_WaitForFirstCompletion":
return self._format_id_list(call, "taskIds", "Wait for")
if call.name == "BackgroundAgents_GetTaskResults":
return self._format_single_id(call, "taskId")
if call.name == "BackgroundAgents_ContinueTask":
return self._format_continue_task(call)
if call.name == "BackgroundAgents_ClearCompletedTask":
return self._format_single_id(call, "taskId")
return None
def _format_start_background_task(self, call: Content) -> str | None:
"""Format StartTask with agent name and description."""
agent_name = get_argument_value(call, "agentName")
description = get_argument_value(call, "description")
if agent_name is None and description is None:
return None
lines: list[str] = []
if agent_name is not None and description is not None:
lines.append(f"\n ├─ Agent: {agent_name}")
lines.append(f'\n └─ "{truncate(description, 80)}"')
elif agent_name is not None:
lines.append(f"\n └─ Agent: {agent_name}")
else:
lines.append(f'\n └─ "{truncate(description, 80)}"') # type: ignore[arg-type]
return "".join(lines)
def _format_id_list(self, call: Content, param_name: str, verb: str) -> str | None:
"""Format a list of task IDs with a verb."""
ids = as_int_list(get_argument_value(call, param_name))
if not ids:
return None
lines: list[str] = []
for i, task_id in enumerate(ids):
connector = "├─" if i < len(ids) - 1 else "└─"
lines.append(f"\n {connector} {verb} #{task_id}")
return "".join(lines)
def _format_single_id(self, call: Content, param_name: str) -> str | None:
"""Format a single task ID in parentheses."""
task_id = get_argument_value(call, param_name)
if isinstance(task_id, int):
return f"(task #{task_id})"
return None
def _format_continue_task(self, call: Content) -> str | None:
"""Format ContinueTask with task ID and optional text."""
task_id = get_argument_value(call, "taskId")
text = get_argument_value(call, "text")
if not isinstance(task_id, int):
return None
if text:
lines = [
f"\n ├─ Task #{task_id}",
f'\n └─ "{truncate(text, 80)}"',
]
return "".join(lines)
return f"\n └─ Task #{task_id}"
class FileMemoryToolFormatter(ToolCallFormatter):
"""Formats FileMemory_* tool calls, showing file names and search patterns
with tree-view corners for save operations.
"""
def can_format(self, call: Content) -> bool:
"""Match FileMemory_* tool calls."""
return call.name is not None and call.name.startswith("FileMemory_")
def format_detail(self, call: Content) -> str | None:
"""Format based on the specific FileMemory operation."""
if call.name == "FileMemory_SaveFile":
return self._format_save_file(call)
if call.name in ("FileMemory_ReadFile", "FileMemory_DeleteFile"):
value = get_argument_value(call, "fileName")
return f"({value})" if value else None
if call.name == "FileMemory_SearchFiles":
return self._format_search_files(call)
return None
def _format_save_file(self, call: Content) -> str | None:
"""Format SaveFile with file name and description indicator."""
file_name = get_argument_value(call, "fileName")
description = get_argument_value(call, "description")
if not file_name:
return None
if description:
return f"\n └─ {file_name} (with description)"
return f"\n └─ {file_name}"
def _format_search_files(self, call: Content) -> str | None:
"""Format SearchFiles with regex pattern and optional file pattern."""
pattern = get_argument_value(call, "regexPattern")
file_pattern = get_argument_value(call, "filePattern")
if not pattern:
return None
if file_pattern:
return f"(/{pattern}/ in {file_pattern})"
return f"(/{pattern}/)"
# endregion
# region Public API Functions
def format_tool_call(formatters: list[ToolCallFormatter], call: Content) -> str:
"""Format a tool call using the first matching formatter from the provided list.
Returns "{toolName} {detail}" when a formatter produces detail,
or just "{toolName}" otherwise.
Args:
formatters: List of formatters to try in order.
call: The function call content to format.
Returns:
Formatted string representation of the tool call.
"""
for formatter in formatters:
if formatter.can_format(call):
detail = formatter.format_detail(call)
tool_name = call.name or "Unknown"
return f"{tool_name} {detail}" if detail is not None else tool_name
return call.name or "Unknown"
def build_default_formatters() -> list[ToolCallFormatter]:
"""Create the default list of tool call formatters.
The FallbackToolFormatter is always last. Users can call this function
and combine the result with their own formatters.
Returns:
A list of all built-in tool call formatters.
"""
return [
TodoToolFormatter(),
ModeToolFormatter(),
BackgroundAgentToolFormatter(),
FileMemoryToolFormatter(),
WebSearchToolFormatter(),
FallbackToolFormatter(),
]
# endregion
@@ -1,87 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Main entry point for the harness console.
Provides the top-level run_agent_async() function that creates and runs
the Textual-based harness console application.
"""
from __future__ import annotations
from typing import TYPE_CHECKING
from .app import HarnessApp
from .observers import build_default_observers
if TYPE_CHECKING:
from agent_framework import Agent, AgentSession
from .commands import CommandHandler
from .observers.base import ConsoleObserver
async def run_agent_async(
agent: Agent,
*,
session: AgentSession | None = None,
observers: list[ConsoleObserver] | None = None,
command_handlers: list[CommandHandler] | None = None,
mode_colors: dict[str, str] | None = None,
initial_mode: str | None = None,
placeholder: str = "Type a message and press Enter...",
title: str = "Harness Console",
max_context_window_tokens: int | None = None,
max_output_tokens: int | None = None,
) -> None:
"""Run the harness console with the given agent.
This is the main entry point for the harness console. Creates a Textual
application with the configured observers and runs it until the user exits.
Args:
agent: The agent to run conversations with.
session: Optional agent session for conversation history.
observers: List of console observers. If None, uses defaults.
command_handlers: List of command handlers. If None, auto-detected from agent.
mode_colors: Mapping of mode names to Rich color strings.
initial_mode: Initial agent mode text.
placeholder: Input placeholder text.
title: Application title.
max_context_window_tokens: Optional max context window size for usage display.
max_output_tokens: Optional max output tokens for usage display.
Example:
.. code-block:: python
from agent_framework import Agent
from agent_framework.openai import OpenAIChatClient
from console import run_agent_async
agent = Agent(
client=OpenAIChatClient(),
instructions="You are helpful.",
)
await run_agent_async(agent)
"""
resolved_observers = observers or build_default_observers()
resolved_mode_colors = mode_colors or {
"plan": "cyan",
"execute": "green",
}
resolved_session = session or agent.create_session()
app = HarnessApp(
agent=agent,
observers=resolved_observers,
session=resolved_session,
mode_colors=resolved_mode_colors,
initial_mode=initial_mode,
placeholder=placeholder,
title=title,
max_context_window_tokens=max_context_window_tokens,
max_output_tokens=max_output_tokens,
command_handlers=command_handlers,
)
await app.run_async()
@@ -1,122 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Console observers for agent streaming lifecycle.
This module provides observers that display events during agent streaming
and collect follow-up actions. All observers use the IUXStateDriver interface
to update the UI.
"""
from __future__ import annotations
from typing import TYPE_CHECKING
from .base import ConsoleObserver
from .error_display import ErrorDisplayObserver
from .planning_output import PlanningOutputObserver
from .reasoning_display import ReasoningDisplayObserver
from .text_output import TextOutputObserver
from .tool_approval import ToolApprovalObserver
from .tool_call_display import ToolCallDisplayObserver
from .usage_display import UsageDisplayObserver
if TYPE_CHECKING:
from agent_framework import Agent
def build_default_observers() -> list[ConsoleObserver]:
"""Build the default set of observers for the harness console.
Returns a standard observer list covering:
- Text output (streaming text display)
- Tool call display (formatted tool invocations)
- Error display (error messages)
- Usage display (token counts)
- Reasoning display (reasoning/thinking blocks)
- Tool approval (user approval for tool calls)
Note: PlanningOutputObserver is NOT included here because it requires
a mode_provider. Use build_observers_with_planning() for agents that
have an AgentModeProvider (i.e. agents created with create_harness_agent).
Returns:
List of default console observers.
"""
return [
TextOutputObserver(),
ToolCallDisplayObserver(),
ErrorDisplayObserver(),
UsageDisplayObserver(),
ReasoningDisplayObserver(),
ToolApprovalObserver(),
]
def build_observers_with_planning(
agent: Agent,
plan_mode_name: str = "plan",
execution_mode_name: str = "execute",
*,
mode_colors: dict[str, str] | None = None,
) -> list[ConsoleObserver]:
"""Build observers with planning support (structured output in plan mode).
Replaces TextOutputObserver with PlanningOutputObserver, which configures
structured JSON output via response_format when in plan mode. This enables
the list picker UI for clarification and approval questions.
Requires that the agent has an AgentModeProvider in its context_providers
(automatically added by create_harness_agent).
Args:
agent: The agent to resolve the AgentModeProvider from.
plan_mode_name: The mode name that represents planning mode.
execution_mode_name: The mode name to switch to on approval.
mode_colors: Optional mapping of mode names to Rich color strings.
Returns:
List of observers with planning support.
Raises:
ValueError: If the agent has no AgentModeProvider.
"""
from agent_framework import AgentModeProvider
mode_provider = next(
(p for p in agent.context_providers if isinstance(p, AgentModeProvider)),
None,
)
if mode_provider is None:
msg = (
"Planning observers require an AgentModeProvider on the agent. "
"Use create_harness_agent() or add AgentModeProvider to context_providers."
)
raise ValueError(msg)
return [
ToolCallDisplayObserver(),
ToolApprovalObserver(),
ErrorDisplayObserver(),
ReasoningDisplayObserver(),
UsageDisplayObserver(),
PlanningOutputObserver(
mode_provider,
plan_mode_name,
execution_mode_name,
mode_colors=mode_colors,
),
]
__all__ = [
"ConsoleObserver",
"ErrorDisplayObserver",
"PlanningOutputObserver",
"ReasoningDisplayObserver",
"TextOutputObserver",
"ToolApprovalObserver",
"ToolCallDisplayObserver",
"UsageDisplayObserver",
"build_default_observers",
"build_observers_with_planning",
]
@@ -1,125 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Base class for console observers.
Observers participate in the agent streaming lifecycle, displaying events
and optionally returning follow-up actions.
"""
from __future__ import annotations
from typing import TYPE_CHECKING, Any
if TYPE_CHECKING:
from agent_framework import Agent, Content, Message
from ..app_state import FollowUpAction
from ..state_driver import IUXStateDriver
class ConsoleObserver:
"""Base class for console observers.
Observers participate in the agent streaming lifecycle, displaying
events (tool calls, errors, reasoning, etc.) and optionally returning
follow-up actions (questions, approval requests).
All methods have default no-op implementations, so subclasses only
override the methods they need.
"""
def configure_run_options(
self,
options: dict[str, Any],
agent: Agent,
session: Any,
) -> None:
"""Configure run options before agent invocation.
Override to set options such as response_format, max_tokens, etc.
Args:
options: Dictionary of chat options to modify.
agent: The AI agent.
session: The agent session.
"""
pass
async def on_response_update(
self,
ux: IUXStateDriver,
update: Message,
agent: Agent,
session: Any,
) -> None:
"""Called for each response update chunk.
Override to inspect update-level metadata or handle provider-specific
events in the raw representation.
Args:
ux: The UX state driver for UI updates.
update: The message update chunk.
agent: The AI agent.
session: The agent session.
"""
pass
async def on_content(
self,
ux: IUXStateDriver,
content: Content,
agent: Agent,
session: Any,
) -> None:
"""Called for each content item in the response.
Override to handle specific content types (function calls, errors, etc.).
Args:
ux: The UX state driver for UI updates.
content: The content item from the response.
agent: The AI agent.
session: The agent session.
"""
pass
async def on_text(
self,
ux: IUXStateDriver,
text: str,
agent: Agent,
session: Any,
) -> None:
"""Called for each text chunk in the response.
Override to accumulate and display streaming text.
Args:
ux: The UX state driver for UI updates.
text: The text chunk.
agent: The AI agent.
session: The agent session.
"""
pass
async def on_stream_complete(
self,
ux: IUXStateDriver,
agent: Agent,
session: Any,
) -> list[FollowUpAction] | None:
"""Called when streaming completes.
Override to return follow-up actions (questions to ask the user,
messages to inject into the next turn, etc.).
Args:
ux: The UX state driver for UI updates.
agent: The AI agent.
session: The agent session.
Returns:
Optional list of follow-up actions to queue, or None.
"""
return None
@@ -1,72 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Error display observer for showing errors."""
from __future__ import annotations
from typing import TYPE_CHECKING, Any
from .base import ConsoleObserver
if TYPE_CHECKING:
from agent_framework import Agent, Content
from ..state_driver import IUXStateDriver
class ErrorDisplayObserver(ConsoleObserver):
"""Displays error content from the agent response.
Shows errors with an ❌ prefix in red to make them easily visible.
"""
async def on_content(
self,
ux: IUXStateDriver,
content: Content,
agent: Agent,
session: Any,
) -> None:
"""Display error content.
Args:
ux: The UX state driver for UI updates.
content: The content item to check for errors.
agent: The AI agent.
session: The agent session.
"""
# Check if this is an error content type
# The exact content type check depends on the agent framework's Content class
if hasattr(content, "type") and content.type == "error":
error_text = self._format_error(content)
ux.append_info_line(error_text, "red")
elif getattr(content, "error", None):
error_text = f"❌ Error: {content.error}" # type: ignore[reportAttributeAccessIssue]
ux.append_info_line(error_text, "red")
def _format_error(self, content: Content) -> str:
"""Format error content for display.
Args:
content: The error content.
Returns:
Formatted error string.
"""
error_text = "❌ Error"
# Try to extract error message
if hasattr(content, "message"):
error_text += f": {content.message}"
elif hasattr(content, "text"):
error_text += f": {content.text}"
# Try to add error code if available
if hasattr(content, "error_code") and content.error_code:
error_text += f" (code: {content.error_code})"
# Try to add details if available
if hasattr(content, "details") and getattr(content, "details", None):
error_text += f" — {content.details}" # type: ignore[reportAttributeAccessIssue]
return error_text
@@ -1,71 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Pydantic models for structured planning output.
These models define the JSON schema that the agent produces when in planning
mode via `response_format`. The schema enables consistent rendering of
clarification questions and approval requests in the console UI.
"""
from __future__ import annotations
from enum import Enum
from pydantic import BaseModel, Field
class PlanningResponseType(str, Enum):
"""Type of planning response from the agent."""
CLARIFICATION = "clarification"
"""The agent needs clarification and presents options for the user to choose from."""
APPROVAL = "approval"
"""The agent is seeking approval to proceed with execution."""
class PlanningQuestion(BaseModel):
"""A single question or item within a PlanningResponse.
For clarification: contains the question text and optional choices.
For approval: contains the plan summary for the user to approve.
"""
message: str = Field(
description=(
"For clarifications, this has the question that needs to be clarified "
"with the user. For approvals, this would contain a summary of the "
"execution plan that the user needs to approve."
),
)
choices: list[str] | None = Field(
default=None,
description=(
"For clarifications, this has a list of options that the user can "
"choose from. null for approvals."
),
)
class PlanningResponse(BaseModel):
"""Structured response from the agent while in planning mode.
Used with structured output (`response_format`) to enable consistent
rendering of clarification questions and approval requests.
"""
type: PlanningResponseType = Field(
description=(
"Use 'clarification' when you need clarification around the user "
"request and you want to present the user with options to choose from. "
"Use 'approval' when you are ready to start execution, but need "
"approval to start executing."
),
)
questions: list[PlanningQuestion] = Field(
description=(
"For clarifications, this has one or more questions to ask the user "
"(each with choices). For approvals, this has exactly one item "
"containing the plan summary for the user to approve."
),
)
@@ -1,242 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Planning output observer for structured agent responses in plan mode.
In planning mode, this observer configures structured JSON output via
response_format, collects streamed text silently, then deserializes the
result as a PlanningResponse to present clarification/approval questions.
In execution mode, text is streamed through directly.
"""
from __future__ import annotations
import json
from typing import TYPE_CHECKING, Any
from rich.markup import escape
from ..app_state import (
ChoiceFollowUpQuestion,
FollowUpAction,
TextFollowUpQuestion,
)
from .base import ConsoleObserver
from .planning_models import PlanningResponse, PlanningResponseType
if TYPE_CHECKING:
from agent_framework import Agent, AgentModeProvider, Message
from ..state_driver import IUXStateDriver
class PlanningOutputObserver(ConsoleObserver):
"""Mode-aware observer that uses structured output in plan mode.
In planning mode:
- Configures response_format to PlanningResponse schema
- Collects streamed text silently
- Deserializes JSON into PlanningResponse
- Builds follow-up questions (clarification or approval)
In execution mode:
- Streams text directly to the UX driver
If JSON parsing fails, falls back to rendering the raw text as regular
output so the user always sees what the agent produced.
"""
def __init__(
self,
mode_provider: AgentModeProvider,
plan_mode_name: str,
execution_mode_name: str,
*,
mode_colors: dict[str, str] | None = None,
) -> None:
"""Initialize the planning output observer.
Args:
mode_provider: The mode provider for reading/switching modes.
plan_mode_name: The mode name that represents planning mode.
execution_mode_name: The mode name to switch to on approval.
mode_colors: Optional mapping of mode names to Rich color strings.
"""
self._mode_provider = mode_provider
self._plan_mode_name = plan_mode_name
self._execution_mode_name = execution_mode_name
self._mode_colors = mode_colors or {}
self._text_collector: list[str] = []
def configure_run_options(
self,
options: dict[str, Any],
agent: Agent,
session: Any,
) -> None:
"""Set response_format to PlanningResponse when in plan mode."""
if self._is_planning_mode(session):
options["response_format"] = PlanningResponse
async def on_text(
self,
ux: IUXStateDriver,
text: str,
agent: Agent,
session: Any,
) -> None:
"""Collect text in plan mode; stream through in execute mode."""
if self._is_planning_mode_from_ux(ux):
self._text_collector.append(text)
else:
ux.write_text(escape(text))
async def on_stream_complete(
self,
ux: IUXStateDriver,
agent: Agent,
session: Any,
) -> list[FollowUpAction] | None:
"""Parse collected text as PlanningResponse and build follow-up actions."""
if not self._is_planning_mode_from_ux(ux):
self._text_collector.clear()
return None
collected_text = "".join(self._text_collector)
self._text_collector.clear()
if not collected_text.strip():
return None
# Attempt to deserialize structured response
try:
planning_response = PlanningResponse.model_validate_json(collected_text)
except (json.JSONDecodeError, ValueError):
# JSON parsing failed — fall back to rendering as regular text
ux.write_text(escape(collected_text))
return None
if planning_response.type == PlanningResponseType.CLARIFICATION:
return self._build_clarification_actions(planning_response)
if planning_response.type == PlanningResponseType.APPROVAL:
if not planning_response.questions:
ux.append_info_line("(approval response had no content)", "yellow")
return None
question = planning_response.questions[0]
return [self._build_approval_action(question, session)]
# Unexpected type — fall back to rendering as regular text
ux.write_text(escape(collected_text))
return None
def _is_planning_mode(self, session: Any) -> bool:
"""Check if session is in planning mode."""
from agent_framework import get_agent_mode
try:
current_mode = get_agent_mode(session)
except (AttributeError, TypeError):
return True # No mode provider → treat as planning
return current_mode.lower() == self._plan_mode_name.lower()
def _is_planning_mode_from_ux(self, ux: IUXStateDriver) -> bool:
"""Check if UX is in planning mode."""
current = ux.current_mode
if current is None:
return True
return current.lower() == self._plan_mode_name.lower()
def _build_clarification_actions(
self,
response: PlanningResponse,
) -> list[FollowUpAction]:
"""Build follow-up questions for clarification."""
actions: list[FollowUpAction] = []
for question in response.questions:
prompt = question.message
cont = self._make_clarification_continuation(prompt)
if question.choices and len(question.choices) > 0:
actions.append(
ChoiceFollowUpQuestion(
prompt=prompt,
choices=question.choices,
allow_custom_text=True,
continuation=cont,
)
)
else:
actions.append(
TextFollowUpQuestion(
prompt=prompt,
continuation=cont,
)
)
return actions
@staticmethod
def _make_clarification_continuation(prompt: str):
"""Create a clarification continuation closure capturing the prompt."""
async def continuation(
answer: str,
ux: IUXStateDriver,
) -> Message | None:
if not answer.strip():
ux.append_info_line(f"🔹 {prompt}\n └─ (no answer)", "dim")
return None
ux.append_info_line(f"🔹 {prompt}\n └─ [green]{answer}[/green]", "dim")
from agent_framework import Message
return Message(role="user", contents=[f"Q: {prompt}\nA: {answer}"])
return continuation
def _build_approval_action(
self,
question: Any,
session: Any,
) -> ChoiceFollowUpQuestion:
"""Build the approval follow-up question."""
approve_option = "Approve and switch to execute mode"
prompt = question.message
async def continuation(
selection: str,
ux: IUXStateDriver,
) -> Message | None:
ux.append_info_line(
f"🔹 {prompt}\n └─ [green]{selection}[/green]",
"dim",
)
if selection == approve_option:
from agent_framework import set_agent_mode
set_agent_mode(session, self._execution_mode_name)
exec_color = self._mode_colors.get(self._execution_mode_name)
ux.set_mode(self._execution_mode_name, exec_color)
ux.append_info_line(
f"âś… Switched to {self._execution_mode_name} mode.",
exec_color,
)
from agent_framework import Message
return Message(role="user", contents=["Approved"])
# Custom freeform input — treat as suggested changes
from agent_framework import Message
return Message(role="user", contents=[selection])
return ChoiceFollowUpQuestion(
prompt=prompt,
choices=[approve_option],
allow_custom_text=True,
continuation=continuation,
)
@@ -1,80 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Reasoning display observer for showing thinking content."""
from __future__ import annotations
from typing import TYPE_CHECKING, Any
from rich.markup import escape
from .base import ConsoleObserver
if TYPE_CHECKING:
from agent_framework import Agent, Content
from ..state_driver import IUXStateDriver
class ReasoningDisplayObserver(ConsoleObserver):
"""Displays reasoning/thinking content from the agent.
Some models (like o1) provide reasoning steps that show their
internal thought process. This observer displays them with a đź’­ prefix
in a dimmed style.
"""
async def on_content(
self,
ux: IUXStateDriver,
content: Content,
agent: Agent,
session: Any,
) -> None:
"""Display reasoning content.
Args:
ux: The UX state driver for UI updates.
content: The content item to check for reasoning.
agent: The AI agent.
session: The agent session.
"""
reasoning_text = self._extract_reasoning(content)
if reasoning_text:
# Display reasoning in dim style to differentiate from main output
ux.append_info_line(f"đź’­ {escape(reasoning_text)}", "dim")
def _extract_reasoning(self, content: Content) -> str | None:
"""Extract reasoning text from content.
Args:
content: The content item to extract reasoning from.
Returns:
The reasoning text, or None if no reasoning is present.
"""
# Check for reasoning content type
if hasattr(content, "type") and content.type in {"text_reasoning", "reasoning"}:
if hasattr(content, "text"):
return content.text
content_attr = getattr(content, "content", None)
if content_attr:
return str(content_attr)
# Check for reasoning attribute
reasoning = getattr(content, "reasoning", None)
if reasoning is not None:
if isinstance(reasoning, str):
return reasoning
if hasattr(reasoning, "text"):
return reasoning.text
# Check for thinking attribute (alternative name)
thinking = getattr(content, "thinking", None)
if thinking is not None:
if isinstance(thinking, str):
return thinking
if hasattr(thinking, "text"):
return thinking.text
return None
@@ -1,59 +0,0 @@
# Copyright (c) Microsoft. All rights reserved.
"""Text output observer for streaming agent text."""
from __future__ import annotations
from typing import TYPE_CHECKING, Any
from rich.markup import escape
from .base import ConsoleObserver
if TYPE_CHECKING:
from agent_framework import Agent
from ..state_driver import IUXStateDriver
class TextOutputObserver(ConsoleObserver):
"""Displays streaming text output from the agent.
Writes text chunks incrementally to the UX state driver as they arrive,
allowing real-time display during streaming.
"""
async def on_text(
self,
ux: IUXStateDriver,
text: str,
agent: Agent,
session: Any,
) -> None:
"""Write each text chunk directly to the UX driver.
Args:
ux: The UX state driver for UI updates.
text: The text chunk to display.
agent: The AI agent.
session: The agent session.
"""
ux.write_text(escape(text))
async def on_stream_complete(
self,
ux: IUXStateDriver,
agent: Agent,
session: Any,
) -> list | None:
"""No-op on stream complete (state managed by UX driver).
Args:
ux: The UX state driver for UI updates.
agent: The AI agent.
session: The agent session.
Returns:
None (no follow-up actions).
"""
return None

Some files were not shown because too many files have changed in this diff Show More