Clarifiied comments and addressed more PR feedbacks.

- pyupgrade: drop forward-reference string annotations in _mcp.py (Python 3.10+ resolves them natively now that MCPTaskOptions is defined before use).

- AGENTS.md: align reconnect description with current behavior. Phase 1 (initial tools/call) does NOT retry on connection loss; raises 'connection lost; task state unknown' instead, so a server that accepted the request but lost the response cannot start the operation twice. Phase 2 (tasks/get / tasks/result) still reconnects once against the same task_id.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

.github/workflows/python-integration-tests.yml

@@ -474,45 +474,6 @@ jobs:
           path: ./python/pytest.xml
           if-no-files-found: ignore
 
-  # GitHub Copilot integration tests
-  python-tests-github-copilot:
-    name: Python Integration Tests - GitHub Copilot
-    runs-on: ubuntu-latest
-    environment: integration
-    timeout-minutes: 60
-    env:
-      COPILOT_GITHUB_TOKEN: ${{ secrets.COPILOT_GITHUB_TOKEN }}
-      GITHUB_COPILOT_TIMEOUT: "120"
-    defaults:
-      run:
-        working-directory: python
-    steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6
-        with:
-          ref: ${{ inputs.checkout-ref }}
-          persist-credentials: false
-      - name: Set up python and install the project
-        id: python-setup
-        uses: ./.github/actions/python-setup
-        with:
-          python-version: ${{ env.UV_PYTHON }}
-          os: ${{ runner.os }}
-      - name: Test with pytest (GitHub Copilot integration)
-        run: >
-          uv run pytest --import-mode=importlib
-          packages/github_copilot/tests
-          -m integration
-          --timeout=120 --session-timeout=900 --timeout_method thread
-          --retries 2 --retry-delay 5
-          --junitxml=pytest.xml
-      - name: Upload test results
-        if: always()
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
-        with:
-          name: test-results-github-copilot
-          path: ./python/pytest.xml
-          if-no-files-found: ignore
-
   # Integration test trend report (aggregates per-job JUnit XML results)
   python-integration-test-report:
     name: Integration Test Report
@@ -529,7 +490,6 @@ jobs:
         python-tests-foundry,
         python-tests-foundry-hosting,
         python-tests-cosmos,
-        python-tests-github-copilot,
       ]
     runs-on: ubuntu-latest
     defaults:
@@ -593,8 +553,7 @@ jobs:
         python-tests-functions,
         python-tests-foundry,
         python-tests-foundry-hosting,
-        python-tests-cosmos,
-        python-tests-github-copilot
+        python-tests-cosmos
       ]
     steps:
       - name: Fail workflow if tests failed

.github/workflows/python-merge-tests.yml

@@ -40,7 +40,6 @@ jobs:
       foundryChanged: ${{ steps.filter.outputs.foundry }}
       foundryHostingChanged: ${{ steps.filter.outputs.foundry_hosting }}
       cosmosChanged: ${{ steps.filter.outputs.cosmos }}
-      githubCopilotChanged: ${{ steps.filter.outputs.github_copilot }}
     steps:
       - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6
       - uses: dorny/paths-filter@d1c1ffe0248fe513906c8e24db8ea791d46f8590 # v3
@@ -86,8 +85,6 @@ jobs:
               - 'python/packages/foundry_hosting/**'
             cosmos:
               - 'python/packages/azure-cosmos/**'
-            github_copilot:
-              - 'python/packages/github_copilot/**'
       # run only if 'python' files were changed
       - name: python tests
         if: steps.filter.outputs.python == 'true'
@@ -661,58 +658,6 @@ jobs:
           path: ./python/pytest.xml
           if-no-files-found: ignore
 
-  # GitHub Copilot integration tests
-  python-tests-github-copilot:
-    name: Python Tests - GitHub Copilot Integration
-    needs: paths-filter
-    if: >
-      github.event_name != 'pull_request' &&
-      needs.paths-filter.outputs.pythonChanges == 'true' &&
-      (github.event_name != 'merge_group' ||
-       needs.paths-filter.outputs.githubCopilotChanged == 'true' ||
-       needs.paths-filter.outputs.coreChanged == 'true')
-    runs-on: ubuntu-latest
-    environment: integration
-    timeout-minutes: 60
-    env:
-      COPILOT_GITHUB_TOKEN: ${{ secrets.COPILOT_GITHUB_TOKEN }}
-      GITHUB_COPILOT_TIMEOUT: "120"
-    defaults:
-      run:
-        working-directory: python
-    steps:
-      - uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6
-      - name: Set up python and install the project
-        id: python-setup
-        uses: ./.github/actions/python-setup
-        with:
-          python-version: ${{ env.UV_PYTHON }}
-          os: ${{ runner.os }}
-      - name: Test with pytest (GitHub Copilot integration)
-        run: >
-          uv run pytest --import-mode=importlib
-          packages/github_copilot/tests
-          -m integration
-          --timeout=120 --session-timeout=900 --timeout_method thread
-          --retries 2 --retry-delay 5
-          --junitxml=pytest.xml
-      - name: Surface failing tests
-        if: always()
-        uses: pmeier/pytest-results-action@20b595761ba9bf89e115e875f8bc863f913bc8ad # v0.7.2
-        with:
-          path: ./python/pytest.xml
-          summary: true
-          display-options: fEX
-          fail-on-empty: false
-          title: GitHub Copilot integration test results
-      - name: Upload test results
-        if: always()
-        uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a # v7
-        with:
-          name: test-results-github-copilot
-          path: ./python/pytest.xml
-          if-no-files-found: ignore
-
   # Integration test trend report (aggregates per-job JUnit XML results)
   python-integration-test-report:
     name: Integration Test Report
@@ -729,7 +674,6 @@ jobs:
         python-tests-foundry,
         python-tests-foundry-hosting,
         python-tests-cosmos,
-        python-tests-github-copilot,
       ]
     runs-on: ubuntu-latest
     defaults:
@@ -791,7 +735,6 @@ jobs:
         python-tests-foundry,
         python-tests-foundry-hosting,
         python-tests-cosmos,
-        python-tests-github-copilot,
       ]
     steps:
       - name: Fail workflow if tests failed

dotnet/Directory.Packages.props

@@ -109,7 +109,7 @@
     <PackageVersion Include="A2A" Version="1.0.0-preview2" />
     <PackageVersion Include="A2A.AspNetCore" Version="1.0.0-preview2" />
     <!-- MCP -->
-    <PackageVersion Include="ModelContextProtocol" Version="1.2.0" />
+    <PackageVersion Include="ModelContextProtocol" Version="1.1.0" />
     <!-- Hyperlight -->
     <PackageVersion Include="Hyperlight.HyperlightSandbox.Api" Version="0.4.0" />
     <PackageVersion Include="Hyperlight.HyperlightSandbox.Guest.Python" Version="0.4.0" />

dotnet/samples/04-hosting/FoundryHostedAgents/responses/Hosted-McpTools/HostedMcpTools.csproj

@@ -13,7 +13,7 @@
   <ItemGroup>
     <PackageReference Include="Azure.AI.Projects" />
     <PackageReference Include="Azure.Identity" />
-    <PackageReference Include="ModelContextProtocol" />
+    <PackageReference Include="ModelContextProtocol" VersionOverride="1.2.0" />
     <PackageReference Include="DotNetEnv" />
   </ItemGroup>
 

dotnet/src/Microsoft.Agents.AI/Skills/File/AgentFileSkill.cs

@@ -49,13 +49,14 @@ public sealed class AgentFileSkill : AgentSkill
     /// <inheritdoc/>
     /// <remarks>
     /// Returns the raw SKILL.md content. When the skill has scripts, a
-    /// <c>&lt;script_schemas&gt;</c> block is appended describing the argument format.
+    /// <c>&lt;scripts&gt;&lt;script name="..."&gt;&lt;parameters_schema&gt;...&lt;/parameters_schema&gt;&lt;/script&gt;&lt;/scripts&gt;</c>
+    /// block is appended with a per-script entry describing the expected argument format.
     /// The result is cached after the first access.
     /// </remarks>
     public override ValueTask<string> GetContentAsync(CancellationToken cancellationToken = default)
     {
         var content = this._content ??= this._scripts is { Count: > 0 }
-            ? this._originalContent + AgentInlineSkillContentBuilder.BuildScriptSchemasBlock(this._scripts)
+            ? this._originalContent + AgentInlineSkillContentBuilder.BuildScriptsBlock(this._scripts)
             : this._originalContent;
         return new(content);
     }

dotnet/src/Microsoft.Agents.AI/Skills/Programmatic/AgentClassSkill.cs

@@ -114,6 +114,7 @@ public abstract class AgentClassSkill<
             this.Frontmatter.Name,
             this.Frontmatter.Description,
             this.Instructions,
+            this.Resources,
             this.Scripts));
     }
 
@@ -146,17 +147,11 @@ public abstract class AgentClassSkill<
     /// Gets the resources associated with this skill, or <see langword="null"/> if none.
     /// </summary>
     /// <remarks>
-    /// <para>
     /// The default implementation returns resources discovered via reflection by scanning
     /// <typeparamref name="TSelf"/> for members annotated with <see cref="AgentSkillResourceAttribute"/>.
     /// This discovery is compatible with Native AOT because <typeparamref name="TSelf"/> is annotated with
     /// <see cref="DynamicallyAccessedMembersAttribute"/>. The result is cached after the first access.
     /// Override this property in derived classes to provide skill-specific resources.
-    /// </para>
-    /// <para>
-    /// Resources are not automatically included in the skill body.
-    /// To enable discovery, reference resources by name in the skill's instructions or in other resources.
-    /// </para>
     /// </remarks>
     public virtual IReadOnlyList<AgentSkillResource>? Resources => this._resources.Value;
 
@@ -164,17 +159,11 @@ public abstract class AgentClassSkill<
     /// Gets the scripts associated with this skill, or <see langword="null"/> if none.
     /// </summary>
     /// <remarks>
-    /// <para>
     /// The default implementation returns scripts discovered via reflection by scanning
     /// <typeparamref name="TSelf"/> for methods annotated with <see cref="AgentSkillScriptAttribute"/>.
     /// This discovery is compatible with Native AOT because <typeparamref name="TSelf"/> is annotated with
     /// <see cref="DynamicallyAccessedMembersAttribute"/>. The result is cached after the first access.
     /// Override this property in derived classes to provide skill-specific scripts.
-    /// </para>
-    /// <para>
-    /// Only script parameter schemas are included in the skill body (as a <c>&lt;script_schemas&gt;</c> block).
-    /// To enable discovery, reference scripts by name in the skill's instructions or in a resource.
-    /// </para>
     /// </remarks>
     public virtual IReadOnlyList<AgentSkillScript>? Scripts => this._scripts.Value;
 
@@ -195,10 +184,6 @@ public abstract class AgentClassSkill<
     /// <summary>
     /// Creates a skill resource backed by a static value.
     /// </summary>
-    /// <remarks>
-    /// Resources are not automatically included in the skill body.
-    /// To enable discovery, reference the resource by name in the skill's instructions or in another resource.
-    /// </remarks>
     /// <param name="name">The resource name.</param>
     /// <param name="value">The static resource value.</param>
     /// <param name="description">An optional description of the resource.</param>
@@ -209,10 +194,6 @@ public abstract class AgentClassSkill<
     /// <summary>
     /// Creates a skill resource backed by a delegate that produces a dynamic value.
     /// </summary>
-    /// <remarks>
-    /// Resources are not automatically included in the skill body.
-    /// To enable discovery, reference the resource by name in the skill's instructions or in another resource.
-    /// </remarks>
     /// <param name="name">The resource name.</param>
     /// <param name="method">A method that produces the resource value when requested.</param>
     /// <param name="description">An optional description of the resource.</param>
@@ -227,10 +208,6 @@ public abstract class AgentClassSkill<
     /// <summary>
     /// Creates a skill script backed by a delegate.
     /// </summary>
-    /// <remarks>
-    /// Only the script's parameter schema is included in the skill body (as a <c>&lt;script_schemas&gt;</c> block).
-    /// To enable discovery, reference the script by name in the skill's instructions or in a resource.
-    /// </remarks>
     /// <param name="name">The script name.</param>
     /// <param name="method">A method to execute when the script is invoked.</param>
     /// <param name="description">An optional description of the script.</param>

dotnet/src/Microsoft.Agents.AI/Skills/Programmatic/AgentInlineSkill.cs

@@ -95,7 +95,7 @@ public sealed class AgentInlineSkill : AgentSkill
     /// <inheritdoc/>
     public override ValueTask<string> GetContentAsync(CancellationToken cancellationToken = default)
     {
-        return new(this._cachedContent ??= AgentInlineSkillContentBuilder.Build(this.Frontmatter.Name, this.Frontmatter.Description, this._instructions, this._scripts));
+        return new(this._cachedContent ??= AgentInlineSkillContentBuilder.Build(this.Frontmatter.Name, this.Frontmatter.Description, this._instructions, this._resources, this._scripts));
     }
 
     /// <inheritdoc/>
@@ -115,10 +115,6 @@ public sealed class AgentInlineSkill : AgentSkill
     /// <summary>
     /// Registers a static resource with this skill.
     /// </summary>
-    /// <remarks>
-    /// Resources are not automatically included in the skill body.
-    /// To enable discovery, reference the resource by name in the skill's instructions or in another resource.
-    /// </remarks>
     /// <param name="name">The resource name.</param>
     /// <param name="value">The static resource value.</param>
     /// <param name="description">An optional description of the resource.</param>
@@ -133,10 +129,6 @@ public sealed class AgentInlineSkill : AgentSkill
     /// Registers a dynamic resource with this skill, backed by a C# delegate.
     /// The delegate's parameters and return type are automatically marshaled via <c>AIFunctionFactory</c>.
     /// </summary>
-    /// <remarks>
-    /// Resources are not automatically included in the skill body.
-    /// To enable discovery, reference the resource by name in the skill's instructions or in another resource.
-    /// </remarks>
     /// <param name="name">The resource name.</param>
     /// <param name="method">A method that produces the resource value when requested.</param>
     /// <param name="description">An optional description of the resource.</param>
@@ -155,10 +147,6 @@ public sealed class AgentInlineSkill : AgentSkill
     /// Registers a script with this skill, backed by a C# delegate.
     /// The delegate's parameters and return type are automatically marshaled via <c>AIFunctionFactory</c>.
     /// </summary>
-    /// <remarks>
-    /// Only the script's parameter schema is included in the skill body (as a <c>&lt;script_schemas&gt;</c> block).
-    /// To enable discovery, reference the script by name in the skill's instructions or in a resource.
-    /// </remarks>
     /// <param name="name">The script name.</param>
     /// <param name="method">A method to execute when the script is invoked.</param>
     /// <param name="description">An optional description of the script.</param>

dotnet/src/Microsoft.Agents.AI/Skills/Programmatic/AgentInlineSkillContentBuilder.cs

@@ -12,17 +12,19 @@ namespace Microsoft.Agents.AI;
 internal static class AgentInlineSkillContentBuilder
 {
     /// <summary>
-    /// Builds the complete skill content containing name, description, instructions, and script parameter schemas.
+    /// Builds the complete skill content containing name, description, instructions, resources, and scripts.
     /// </summary>
     /// <param name="name">The skill name.</param>
     /// <param name="description">The skill description.</param>
     /// <param name="instructions">The raw instructions text.</param>
+    /// <param name="resources">Optional resources associated with the skill.</param>
     /// <param name="scripts">Optional scripts associated with the skill.</param>
     /// <returns>An XML-structured content string.</returns>
     public static string Build(
         string name,
         string description,
         string instructions,
+        IReadOnlyList<AgentSkillResource>? resources,
         IReadOnlyList<AgentSkillScript>? scripts)
     {
         _ = Throw.IfNullOrWhitespace(name);
@@ -37,24 +39,41 @@ internal static class AgentInlineSkillContentBuilder
         .Append(EscapeXmlString(instructions))
         .Append("\n</instructions>");
 
+        if (resources is { Count: > 0 })
+        {
+            sb.Append("\n\n<resources>\n");
+            foreach (var resource in resources)
+            {
+                if (resource.Description is not null)
+                {
+                    sb.Append($"  <resource name=\"{EscapeXmlString(resource.Name)}\" description=\"{EscapeXmlString(resource.Description)}\"/>\n");
+                }
+                else
+                {
+                    sb.Append($"  <resource name=\"{EscapeXmlString(resource.Name)}\"/>\n");
+                }
+            }
+
+            sb.Append("</resources>");
+        }
+
         if (scripts is { Count: > 0 })
         {
             sb.Append('\n');
-            sb.Append(BuildScriptSchemasBlock(scripts));
+            sb.Append(BuildScriptsBlock(scripts));
         }
 
         return sb.ToString();
     }
 
     /// <summary>
-    /// Builds a <c>&lt;script_schemas&gt;...&lt;/script_schemas&gt;</c> XML block for the given scripts.
-    /// Each script is emitted as a <c>&lt;schema script="..."&gt;</c> element containing only
-    /// the parameter schema. This block serves as a reference for the model to know how to
-    /// format arguments when calling scripts, not as a discovery mechanism.
+    /// Builds a <c>&lt;scripts&gt;...&lt;/scripts&gt;</c> XML block for the given scripts.
+    /// Each script is emitted as a <c>&lt;script name="..."&gt;</c> element with optional
+    /// <c>description</c> attribute and <c>&lt;parameters_schema&gt;</c> child element.
     /// </summary>
     /// <param name="scripts">The scripts to include in the block.</param>
-    /// <returns>An XML string starting with <c>\n&lt;script_schemas&gt;</c>, or an empty string if the list is empty.</returns>
-    public static string BuildScriptSchemasBlock(IReadOnlyList<AgentSkillScript> scripts)
+    /// <returns>An XML string starting with <c>\n&lt;scripts&gt;</c>, or an empty string if the list is empty.</returns>
+    public static string BuildScriptsBlock(IReadOnlyList<AgentSkillScript> scripts)
     {
         _ = Throw.IfNull(scripts);
 
@@ -64,23 +83,32 @@ internal static class AgentInlineSkillContentBuilder
         }
 
         var sb = new StringBuilder();
-        sb.Append("\n<script_schemas>\n");
+        sb.Append("\n<scripts>\n");
 
         foreach (var script in scripts)
         {
             var parametersSchema = script.ParametersSchema;
 
-            if (parametersSchema is null)
+            if (script.Description is null && parametersSchema is null)
             {
-                sb.Append($"  <schema script=\"{EscapeXmlString(script.Name)}\"/>\n");
+                sb.Append($"  <script name=\"{EscapeXmlString(script.Name)}\"/>\n");
             }
             else
             {
-                sb.Append($"  <schema script=\"{EscapeXmlString(script.Name)}\">{EscapeXmlString(parametersSchema.Value.GetRawText(), preserveQuotes: true)}</schema>\n");
+                sb.Append(script.Description is not null
+                    ? $"  <script name=\"{EscapeXmlString(script.Name)}\" description=\"{EscapeXmlString(script.Description)}\">\n"
+                    : $"  <script name=\"{EscapeXmlString(script.Name)}\">\n");
+
+                if (parametersSchema is not null)
+                {
+                    sb.Append($"    <parameters_schema>{EscapeXmlString(parametersSchema.Value.GetRawText(), preserveQuotes: true)}</parameters_schema>\n");
+                }
+
+                sb.Append("  </script>\n");
             }
         }
 
-        sb.Append("</script_schemas>");
+        sb.Append("</scripts>");
 
         return sb.ToString();
     }

dotnet/tests/Microsoft.Agents.AI.UnitTests/AgentSkills/AgentClassSkillTests.cs

@@ -51,8 +51,9 @@ public sealed class AgentClassSkillTests
         // Act & Assert — Content is cached
         Assert.Same(await skill.GetContentAsync(), await skill.GetContentAsync());
 
-        // Act & Assert — Content includes parameter schema from typed script (with preserved quotes)
-        Assert.Contains("\"value\"", await skill.GetContentAsync());
+        // Act & Assert — Content includes parameter schema from typed script
+        Assert.Contains("parameters_schema", await skill.GetContentAsync());
+        Assert.Contains("value", await skill.GetContentAsync());
     }
 
     [Fact]
@@ -382,9 +383,10 @@ public sealed class AgentClassSkillTests
         // Arrange
         var skill = new AttributedFullSkill();
 
-        // Act & Assert — Content no longer includes resources in body; scripts are in script_schemas
-        Assert.DoesNotContain("<resources>", await skill.GetContentAsync());
-        Assert.Contains("<script_schemas>", await skill.GetContentAsync());
+        // Act & Assert — Content includes reflected resources and scripts
+        Assert.Contains("<resources>", await skill.GetContentAsync());
+        Assert.Contains("conversion-table", await skill.GetContentAsync());
+        Assert.Contains("<scripts>", await skill.GetContentAsync());
         Assert.Contains("convert", await skill.GetContentAsync());
 
         // Act & Assert — discovered members are cached
@@ -502,7 +504,7 @@ public sealed class AgentClassSkillTests
     }
 
     [Fact]
-    public async Task Content_DoesNotRenderResources_InBodyAsync()
+    public async Task Content_IncludesDescription_ForReflectedResourcesAsync()
     {
         // Arrange
         var skill = new AttributedResourcePropertiesSkill();
@@ -510,8 +512,8 @@ public sealed class AgentClassSkillTests
         // Act
         var content = await skill.GetContentAsync();
 
-        // Assert — resources are no longer rendered in body content
-        Assert.DoesNotContain("<resources>", content);
+        // Assert — descriptions from [Description] attribute appear in synthesized content
+        Assert.Contains("Some important data.", content);
     }
 
     [Fact]

dotnet/tests/Microsoft.Agents.AI.UnitTests/AgentSkills/AgentFileSkillScriptTests.cs

@@ -122,10 +122,11 @@ public sealed class AgentFileSkillScriptTests
 
         // Assert — content starts with original and appends per-script entries
         Assert.StartsWith("Original content", content);
-        Assert.Contains("<script_schemas>", content);
-        Assert.Contains("<schema script=\"build\">", content);
-        Assert.Contains("<schema script=\"deploy\">", content);
-        Assert.Contains("</script_schemas>", content);
+        Assert.Contains("<scripts>", content);
+        Assert.Contains("<script name=\"build\">", content);
+        Assert.Contains("<script name=\"deploy\">", content);
+        Assert.Contains("<parameters_schema>", content);
+        Assert.Contains("</scripts>", content);
     }
 
     [Fact]

dotnet/tests/Microsoft.Agents.AI.UnitTests/AgentSkills/AgentInlineSkillTests.cs

@@ -149,7 +149,7 @@ public sealed class AgentInlineSkillTests
     }
 
     [Fact]
-    public async Task Content_DoesNotIncludeResourcesInBodyAsync()
+    public async Task Content_IncludesResourcesAddedBeforeFirstAccessAsync()
     {
         // Arrange
         var skill = new AgentInlineSkill("my-skill", "A valid skill.", "Instructions.");
@@ -158,12 +158,13 @@ public sealed class AgentInlineSkillTests
         // Act
         var content = await skill.GetContentAsync();
 
-        // Assert — resources are no longer rendered in the body; they're accessed via GetResourceAsync
-        Assert.DoesNotContain("<resources>", content);
+        // Assert
+        Assert.Contains("<resources>", content);
+        Assert.Contains("config", content);
     }
 
     [Fact]
-    public async Task Content_DoesNotIncludeDelegateResourcesInBodyAsync()
+    public async Task Content_IncludesDelegateResourcesAddedBeforeFirstAccessAsync()
     {
         // Arrange
         var skill = new AgentInlineSkill("my-skill", "A valid skill.", "Instructions.");
@@ -172,8 +173,9 @@ public sealed class AgentInlineSkillTests
         // Act
         var content = await skill.GetContentAsync();
 
-        // Assert — resources are no longer rendered in the body
-        Assert.DoesNotContain("<resources>", content);
+        // Assert
+        Assert.Contains("<resources>", content);
+        Assert.Contains("dynamic", content);
     }
 
     [Fact]
@@ -187,7 +189,7 @@ public sealed class AgentInlineSkillTests
         var content = await skill.GetContentAsync();
 
         // Assert
-        Assert.Contains("<script_schemas>", content);
+        Assert.Contains("<scripts>", content);
         Assert.Contains("run", content);
     }
 
@@ -207,7 +209,7 @@ public sealed class AgentInlineSkillTests
     }
 
     [Fact]
-    public async Task Content_IncludesScriptSchemasAddedBeforeFirstAccessAsync()
+    public async Task Content_IncludesResourcesAndScriptsAddedBeforeFirstAccessAsync()
     {
         // Arrange
         var skill = new AgentInlineSkill("my-skill", "A valid skill.", "Instructions.");
@@ -218,8 +220,9 @@ public sealed class AgentInlineSkillTests
         var content = await skill.GetContentAsync();
 
         // Assert
-        Assert.DoesNotContain("<resources>", content);
-        Assert.Contains("<script_schemas>", content);
+        Assert.Contains("<resources>", content);
+        Assert.Contains("r1", content);
+        Assert.Contains("<scripts>", content);
         Assert.Contains("s1", content);
     }
 
@@ -233,9 +236,8 @@ public sealed class AgentInlineSkillTests
         // Act
         var content = await skill.GetContentAsync();
 
-        // Assert — JSON schema should be present inside <schema> element (no extra wrapper) with preserved quotes
-        Assert.Contains("<schema script=\"search\">", content);
-        Assert.Contains("\"query\"", content);
+        // Assert — JSON schema should be present and XML content chars escaped
+        Assert.Contains("parameters_schema", content);
         Assert.DoesNotContain("<![CDATA[", content);
     }
 
@@ -427,7 +429,7 @@ public sealed class AgentInlineSkillTests
 
         // Assert
         Assert.DoesNotContain("<resources>", content);
-        Assert.DoesNotContain("<script_schemas>", content);
+        Assert.DoesNotContain("<scripts>", content);
     }
 
     [Fact]
@@ -461,7 +463,7 @@ public sealed class AgentInlineSkillTests
     }
 
     [Fact]
-    public async Task Content_ScriptWithDescription_DoesNotEmitDescriptionAttributeAsync()
+    public async Task Content_ScriptWithDescription_IncludesDescriptionAttributeAsync()
     {
         // Arrange
         var skill = new AgentInlineSkill("my-skill", "A valid skill.", "Instructions.");
@@ -470,10 +472,8 @@ public sealed class AgentInlineSkillTests
         // Act
         var content = await skill.GetContentAsync();
 
-        // Assert — description is no longer emitted in the script_schemas block;
-        // the block only contains parameter schemas for calling scripts.
-        Assert.Contains("<schema script=\"my-script\"", content);
-        Assert.DoesNotContain("description=\"Runs something.\"", content);
+        // Assert
+        Assert.Contains("description=\"Runs something.\"", content);
     }
 
     [Fact]
@@ -492,7 +492,7 @@ public sealed class AgentInlineSkillTests
     }
 
     [Fact]
-    public async Task Content_ResourceWithDescription_NotRenderedInBodyAsync()
+    public async Task Content_ResourceWithDescription_IncludesDescriptionAttributeAsync()
     {
         // Arrange
         var skill = new AgentInlineSkill("my-skill", "A valid skill.", "Instructions.");
@@ -502,10 +502,9 @@ public sealed class AgentInlineSkillTests
         // Act
         var content = await skill.GetContentAsync();
 
-        // Assert — resources are no longer rendered in the body
-        Assert.DoesNotContain("<resources>", content);
-        Assert.DoesNotContain("with-desc", content);
-        Assert.DoesNotContain("no-desc", content);
+        // Assert
+        Assert.Contains("description=\"A described resource.\"", content);
+        Assert.DoesNotContain("no-desc\" description", content);
     }
 
     [Fact]

python/CHANGELOG.md

@@ -7,43 +7,6 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
-## [1.8.0] - 2026-06-04
-
-### Added
-- **agent-framework-core**: Add MCP-based skills discovery (`McpSkillsSource`) ([#6169](https://github.com/microsoft/agent-framework/pull/6169))
-- **agent-framework-core**: Progressive tool exposure via `FunctionInvocationContext` ([#6233](https://github.com/microsoft/agent-framework/pull/6233))
-- **agent-framework-core**: Add background agent support to harness agent ([#6155](https://github.com/microsoft/agent-framework/pull/6155))
-- **agent-framework-core**: Add `AgentFileStore` and `FileAccessProvider` for file access operations ([#6099](https://github.com/microsoft/agent-framework/pull/6099))
-- **agent-framework-core**: Coalesce code interpreter history chunks ([#5801](https://github.com/microsoft/agent-framework/pull/5801))
-- **agent-framework-core**: Run sync tools off the event loop ([#5773](https://github.com/microsoft/agent-framework/pull/5773))
-- **agent-framework-bedrock**: Implement native structured output support via Converse API ([#6052](https://github.com/microsoft/agent-framework/pull/6052))
-- **agent-framework-foundry**: Add Foundry Adaptive Evals integration for rubric-generation ([#6101](https://github.com/microsoft/agent-framework/pull/6101))
-- **agent-framework-foundry**: Add `timeout` parameter to `FoundryAgent` to fix `ConnectTimeout` on multi-turn conversations ([#6263](https://github.com/microsoft/agent-framework/pull/6263))
-- **agent-framework-mistral**: Add Mistral AI embedding client package ([#5480](https://github.com/microsoft/agent-framework/pull/5480))
-- **agent-framework-a2a**: Expose `supported_protocol_bindings` as configurable parameter ([#6098](https://github.com/microsoft/agent-framework/pull/6098))
-- **agent-framework-a2a**: Set `message_id` on `AgentResponseUpdate` for message-bearing paths ([#6163](https://github.com/microsoft/agent-framework/pull/6163))
-- **agent-framework-foundry-hosting**: Persist hosted MCP call/results as canonical `mcp_call` output ([#6070](https://github.com/microsoft/agent-framework/pull/6070))
-
-### Changed
-- **agent-framework-github-copilot**: [BREAKING] Upgrade `github-copilot-sdk` to v1.0.0 (stable) ([#6292](https://github.com/microsoft/agent-framework/pull/6292))
-- **agent-framework-core**: [BREAKING — experimental] Refactor Skill API to async resource and script lookup ([#6135](https://github.com/microsoft/agent-framework/pull/6135))
-- **agent-framework-github-copilot**: Promote to release candidate (`1.0.0rc1`)
-- **agent-framework-declarative**: Promote to release candidate (`1.0.0rc1`) ([#6256](https://github.com/microsoft/agent-framework/pull/6256))
-
-### Fixed
-- **agent-framework-core**: Fix compaction message-id collisions and tool-loop summary persistence ([#6299](https://github.com/microsoft/agent-framework/pull/6299))
-- **agent-framework-core**: Fix observability unsafe serialization of function-call arguments containing dataclass/framework objects ([#6026](https://github.com/microsoft/agent-framework/pull/6026))
-- **agent-framework-core**: Consolidate MCP reliability fixes ([#6145](https://github.com/microsoft/agent-framework/pull/6145))
-- **agent-framework-core**: Backfill chat span request model if unknown and response model is available ([#6160](https://github.com/microsoft/agent-framework/pull/6160))
-- **agent-framework-anthropic**: Skip orphan anthropic thinking signatures ([#5784](https://github.com/microsoft/agent-framework/pull/5784))
-- **agent-framework-foundry**: Fix `FoundryAgent` stripping model from `PromptAgent` requests ([#5526](https://github.com/microsoft/agent-framework/pull/5526))
-- **agent-framework-foundry-hosting**: Fix toolbox consent flow in hosted agent ([#6249](https://github.com/microsoft/agent-framework/pull/6249))
-- **agent-framework-foundry-hosting**: Drop hosted MCP calls when reasoning is stripped ([#6210](https://github.com/microsoft/agent-framework/pull/6210))
-- **agent-framework-openai**: Fix OTLP HTTP base-endpoint losing `/v1/{signal}` auto-append ([#5913](https://github.com/microsoft/agent-framework/pull/5913))
-- **agent-framework-openai**: Drop hosted MCP calls when reasoning is stripped ([#6210](https://github.com/microsoft/agent-framework/pull/6210))
-- **agent-framework-orchestrations**: Fix spurious Magentic custom manager warning ([#6261](https://github.com/microsoft/agent-framework/pull/6261))
-- **agent-framework-azurefunctions**: Fix integration test worker crashes on Py3.13 ([#4260](https://github.com/microsoft/agent-framework/pull/4260))
-
 ## [1.7.0] - 2026-05-28
 
 ### Added
@@ -1169,8 +1132,7 @@ 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-1.8.0...HEAD
-[1.8.0]: https://github.com/microsoft/agent-framework/compare/python-1.7.0...python-1.8.0
+[Unreleased]: https://github.com/microsoft/agent-framework/compare/python-1.7.0...HEAD
 [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
 [1.5.0]: https://github.com/microsoft/agent-framework/compare/python-1.4.0...python-1.5.0

python/PACKAGE_STATUS.md

@@ -33,7 +33,7 @@ Status is grouped into these buckets:
 | `agent-framework-foundry` | `python/packages/foundry` | `released` |
 | `agent-framework-foundry-local` | `python/packages/foundry_local` | `beta` |
 | `agent-framework-gemini` | `python/packages/gemini` | `alpha` |
-| `agent-framework-github-copilot` | `python/packages/github_copilot` | `rc` |
+| `agent-framework-github-copilot` | `python/packages/github_copilot` | `beta` |
 | `agent-framework-hyperlight` | `python/packages/hyperlight` | `beta` |
 | `agent-framework-lab` | `python/packages/lab` | `beta` |
 | `agent-framework-mem0` | `python/packages/mem0` | `beta` |

python/packages/a2a/pyproject.toml

@@ -4,7 +4,7 @@ description = "A2A integration for Microsoft Agent Framework."
 authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
 readme = "README.md"
 requires-python = ">=3.10"
-version = "1.0.0b260604"
+version = "1.0.0b260528"
 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>=1.8.0,<2",
+    "agent-framework-core>=1.7.0,<2",
     "a2a-sdk>=1.0.0,<2",
 ]
 

python/packages/anthropic/pyproject.toml

@@ -4,7 +4,7 @@ description = "Anthropic integration for Microsoft Agent Framework."
 authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
 readme = "README.md"
 requires-python = ">=3.10"
-version = "1.0.0b260604"
+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>=1.8.0,<2",
+    "agent-framework-core>=1.6.0,<2",
     "anthropic>=0.80.0,<0.80.1",
 ]
 

python/packages/azurefunctions/pyproject.toml

@@ -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 = "1.0.0b260604"
+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"
@@ -22,8 +22,8 @@ classifiers = [
   "Typing :: Typed",
 ]
 dependencies = [
-    "agent-framework-core>=1.8.0,<2",
-    "agent-framework-durabletask>=1.0.0b260604,<2",
+    "agent-framework-core>=1.6.0,<2",
+    "agent-framework-durabletask>=1.0.0b260521,<2",
     "azure-functions>=1.24.0,<2",
     "azure-functions-durable>=1.3.1,<2",
 ]

python/packages/bedrock/agent_framework_bedrock/_chat_client.py

@@ -795,7 +795,10 @@ class BedrockChatClient(
             schema = copy.deepcopy(schema_src)
         else:
             if not isinstance(response_format, type) or not issubclass(response_format, BaseModel):
-                raise TypeError("response_format must be None, a dict JSON schema, or a Pydantic BaseModel subclass.")
+                raise TypeError(
+                    "response_format must be None, a dict JSON schema, "
+                    "or a Pydantic BaseModel subclass."
+                )
             # response_format is a Pydantic model class
             schema = response_format.model_json_schema()
             name = response_format.__name__
@@ -814,7 +817,9 @@ class BedrockChatClient(
         return {
             "textFormat": {
                 "type": "json_schema",
-                "structure": {"jsonSchema": json_schema},
+                "structure": {
+                    "jsonSchema": json_schema
+                },
             }
         }
 
@@ -835,7 +840,9 @@ class BedrockChatClient(
                 if node_id in visited:
                     return
                 visited.add(node_id)
-                if node.get("type") == "object" or ("properties" in node and "type" not in node):
+                if node.get("type") == "object" or (
+                    "properties" in node and "type" not in node
+                ):
                     existing = node.get("additionalProperties")
                     if existing is None or existing is True:
                         node["additionalProperties"] = False

python/packages/bedrock/pyproject.toml

@@ -4,7 +4,7 @@ description = "Amazon Bedrock integration for Microsoft Agent Framework."
 authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
 readme = "README.md"
 requires-python = ">=3.10"
-version = "1.0.0b260604"
+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>=1.8.0,<2",
+    "agent-framework-core>=1.6.0,<2",
     "boto3>=1.35.0,<2.0.0",
     "botocore>=1.35.0,<2.0.0",
 ]

python/packages/bedrock/tests/test_bedrock_structured_output.py

@@ -238,7 +238,6 @@ async def test_chat_response_value_populated_streaming() -> None:
 
 async def test_unsupported_model_validation_exception() -> None:
     """When a model doesn't support outputConfig, a clear error should be raised."""
-
     class _FailingStubBedrockRuntime:
         def converse(self, **kwargs: Any) -> dict[str, Any]:
             # Simulate botocore ClientError for ValidationException

python/packages/core/AGENTS.md

@@ -76,6 +76,19 @@ agent_framework/
 - **`SkillScriptRunner`** - Protocol for file-based script execution. Any callable matching `(skill, script, args) -> Any` satisfies it. Code-defined scripts do not use a runner.
 - **`SkillsProvider`** - Context provider (extends `ContextProvider`) that discovers file-based skills from `SKILL.md` files and/or accepts code-defined `Skill` instances. Follows progressive disclosure: advertise → load → read resources / run scripts.
 
+### Model Context Protocol (`_mcp.py`)
+
+- **`MCPTool`** - Base wrapper that owns the MCP `ClientSession` and exposes the remote server's tools as `FunctionTool`s.
+- **`MCPStdioTool`** / **`MCPStreamableHTTPTool`** / **`MCPWebsocketTool`** - Transport-specific subclasses.
+- **`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.
+  - `max_task_wait: timedelta | None` — client-side deadline for the whole post-create lifecycle (poll + result fetch). When exceeded, raises `ToolExecutionException` and fires a best-effort `tasks/cancel`. `None` (default) means no client-side bound. Bounds sleeps, sends, AND reconnects via `asyncio.wait_for`.
+- **Permissive fallback**: servers that ignore the augmentation (return `CallToolResult` directly) or reject the unknown `task` field with `METHOD_NOT_FOUND` / `INVALID_PARAMS` fall back to the plain `session.call_tool(...)` path so legacy servers keep working. An unparseable success response (server accepted the augmented call but returned a payload that is neither `CreateTaskResult` nor `CallToolResult`) **does not** fall back — it raises `ToolExecutionException` to avoid double-executing a side-effecting tool.
+- **Submit-vs-track reconnect policy**: a dropped connection before a `task_id` is known raises `ToolExecutionException("connection lost; task state unknown")` without re-issuing the augmented `tools/call`, so a server that accepted the request but lost the response cannot be made to start the same operation twice; once a `task_id` exists, `tasks/get` / `tasks/result` reconnect once and retry against the same id (a shared `_send_with_one_reconnect` helper).
+- **Cancel-on-abandonment vs terminal failure**: any path where the remote task may still be running (max-wait exceeded, hard `McpError` in poll, malformed `tasks/get`, second connection loss in poll/fetch, reconnect failure) fires best-effort `tasks/cancel` before raising. Terminal failures (`failed`/`cancelled`/`input_required` server-side, `completed+isError`, malformed `tasks/result` after server completed) do **not** cancel — the server is already done. `_MCPTaskAbandoned` is the private marker distinguishing the two.
+- **Transient poll retry**: a slow `tasks/get` that surfaces as `McpError(code=408 REQUEST_TIMEOUT)` is retried (bounded by `max_task_wait`). All other non-connection `McpError`s during poll are treated as abandonment. `tasks/result` does not get transient retry — the server has already completed, so a slow payload fetch is anomalous.
+
 ### File Access Harness (`_harness/_file_access.py`)
 
 - **`AgentFileStore`** - Abstract async store backing the file-access harness. Implementations expose `write_file`, `read_file`, `delete_file`, `list_files`, `file_exists`, `search_files`, and `create_directory` over forward-slash relative paths.

python/packages/core/agent_framework/__init__.py

@@ -124,7 +124,7 @@ from ._harness._todo import (
     TodoSessionStore,
     TodoStore,
 )
-from ._mcp import MCPStdioTool, MCPStreamableHTTPTool, MCPWebsocketTool
+from ._mcp import MCPStdioTool, MCPStreamableHTTPTool, MCPTaskOptions, MCPWebsocketTool
 from ._middleware import (
     AgentContext,
     AgentMiddleware,
@@ -444,12 +444,13 @@ __all__ = [
     "InlineSkillResource",
     "InlineSkillScript",
     "LocalEvaluator",
-    "MCPStdioTool",
-    "MCPStreamableHTTPTool",
-    "MCPWebsocketTool",
     "MCPSkill",
     "MCPSkillResource",
     "MCPSkillsSource",
+    "MCPStdioTool",
+    "MCPStreamableHTTPTool",
+    "MCPTaskOptions",
+    "MCPWebsocketTool",
     "MemoryContextProvider",
     "MemoryFileStore",
     "MemoryIndexEntry",

python/packages/core/agent_framework/_clients.py

@@ -380,15 +380,8 @@ class BaseChatClient(SerializationMixin, ABC, Generic[OptionsCoT]):
             return prepared_messages
         from ._compaction import apply_compaction
 
-        # Compact the caller's list in place when possible. A compaction operation has
-        # two halves: exclusion flags (mutated on shared Message objects) and inserted
-        # summary messages. Operating on the original list keeps both halves on the list
-        # the function-invocation tool loop reuses across iterations; otherwise inserted
-        # summaries would be lost on a throwaway copy while exclusions persisted, silently
-        # dropping older groups (issue #4991).
-        working_messages = messages if isinstance(messages, list) else prepared_messages
         return await apply_compaction(
-            working_messages,
+            prepared_messages,
             strategy=compaction_strategy,
             tokenizer=tokenizer,
         )

python/packages/core/agent_framework/_compaction.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 
 import json
 import logging
-from collections.abc import Iterable, Mapping, Sequence
+from collections.abc import Mapping, Sequence
 from typing import (
     TYPE_CHECKING,
     Any,
@@ -92,23 +92,10 @@ def _is_reasoning_only_assistant(message: Message) -> bool:
     return all(content.type == "text_reasoning" for content in message.contents)
 
 
-def _ensure_message_ids(
-    messages: list[Message], *, id_offset: int = 0, reserved_ids: Iterable[str] | None = None
-) -> None:
-    existing_ids: set[str] = set(reserved_ids) if reserved_ids is not None else set()
-    existing_ids.update(message.message_id for message in messages if message.message_id)
+def _ensure_message_ids(messages: list[Message]) -> None:
     for index, message in enumerate(messages):
-        if message.message_id:
-            continue
-        candidate = f"msg_{id_offset + index}"
-        if candidate in existing_ids:
-            counter = id_offset + len(messages)
-            candidate = f"msg_{counter}"
-            while candidate in existing_ids:
-                counter += 1
-                candidate = f"msg_{counter}"
-        message.message_id = candidate
-        existing_ids.add(candidate)
+        if not message.message_id:
+            message.message_id = f"msg_{index}"
 
 
 def _group_id_for(message: Message, group_index: int) -> str:
@@ -117,27 +104,14 @@ def _group_id_for(message: Message, group_index: int) -> str:
     return f"group_index_{group_index}"
 
 
-def group_messages(
-    messages: list[Message], *, id_offset: int = 0, reserved_ids: Iterable[str] | None = None
-) -> list[dict[str, Any]]:
+def group_messages(messages: list[Message]) -> list[dict[str, Any]]:
     """Compute group spans and metadata for annotation.
 
-    Args:
-        messages: The messages (or a slice of them) to group.
-
-    Keyword Args:
-        id_offset: Absolute starting index used when auto-assigning ``message_id``
-            values, so incremental annotation of a list slice produces ids that
-            stay unique across the full list.
-        reserved_ids: Message ids that already exist outside ``messages`` (for
-            example in a preserved prefix). Auto-assigned ids are guaranteed not
-            to collide with these, preventing duplicate ids across the full list.
-
     Returns:
         Ordered list of lightweight span dicts with keys:
         ``group_id``, ``kind``, ``start_index``, ``end_index``, ``has_reasoning``.
     """
-    _ensure_message_ids(messages, id_offset=id_offset, reserved_ids=reserved_ids)
+    _ensure_message_ids(messages)
     spans: list[dict[str, Any]] = []
     i = 0
     group_index = 0
@@ -465,8 +439,7 @@ def annotate_message_groups(
         if previous_group_index is not None:
             group_index_offset = previous_group_index + 1
 
-    reserved_ids = {message.message_id for message in messages[:start_index] if message.message_id}
-    spans = group_messages(messages[start_index:], id_offset=start_index, reserved_ids=reserved_ids)
+    spans = group_messages(messages[start_index:])
     for span_index, span in enumerate(spans):
         group_id = str(span["group_id"])
         kind = _coerce_group_kind(span["kind"])

python/packages/core/agent_framework/_feature_stage.py

@@ -58,6 +58,7 @@ class ExperimentalFeature(str, Enum):
     FOUNDRY_PREVIEW_TOOLS = "FOUNDRY_PREVIEW_TOOLS"
     FUNCTIONAL_WORKFLOWS = "FUNCTIONAL_WORKFLOWS"
     HARNESS = "HARNESS"
+    MCP_LONG_RUNNING_TASKS = "MCP_LONG_RUNNING_TASKS"
     MCP_SKILLS = "MCP_SKILLS"
     PROGRESSIVE_TOOLS = "PROGRESSIVE_TOOLS"
     SKILLS = "SKILLS"

python/packages/core/agent_framework/_harness/_background_agents.py

@@ -349,8 +349,6 @@ class BackgroundAgentsProvider(ContextProvider):
             _save_provider_state(session, provider_state, source_id=source_id)
             return f"Background task {task_id} started on agent '{agent_name}'."
 
-        background_agents_start_task._invoke_sync_on_event_loop = True  # pyright: ignore[reportPrivateUsage]
-
         @tool(name="background_agents_wait_for_first_completion", approval_mode="never_require")
         async def background_agents_wait_for_first_completion(task_ids: list[int]) -> str:
             """Block until the first of the specified background tasks completes. Returns the completed task's ID."""
@@ -473,8 +471,6 @@ class BackgroundAgentsProvider(ContextProvider):
             _save_provider_state(session, provider_state, source_id=source_id)
             return f"Task {task_id} continued with new input."
 
-        background_agents_continue_task._invoke_sync_on_event_loop = True  # pyright: ignore[reportPrivateUsage]
-
         @tool(name="background_agents_clear_completed_task", approval_mode="never_require")
         def background_agents_clear_completed_task(task_id: int) -> str:
             """Remove a completed or failed task and release its session to free memory."""

python/packages/core/agent_framework/_mcp.py

@@ -4,6 +4,7 @@ from __future__ import annotations
 
 import asyncio
 import base64
+import contextlib
 import contextvars
 import json
 import logging
@@ -12,12 +13,14 @@ import sys
 from abc import abstractmethod
 from collections.abc import Callable, Collection, Coroutine, Mapping, Sequence
 from contextlib import AsyncExitStack, _AsyncGeneratorContextManager  # type: ignore
+from dataclasses import dataclass
 from datetime import timedelta
 from functools import partial
 from typing import TYPE_CHECKING, Any, Literal, TypedDict, cast
 
 from opentelemetry import propagate
 
+from ._feature_stage import ExperimentalFeature, experimental
 from ._tools import FunctionTool
 from ._types import (
     ChatOptions,
@@ -149,6 +152,73 @@ def _url_origin(url: Any) -> tuple[str, str, int | None]:
     return (url.scheme, url.host or "", port)
 
 
+# Internal polling bounds for MCP long-running tasks. Not user-tunable today;
+# promote to MCPTaskOptions if a concrete need arises.
+_MCP_TASK_MIN_POLL_INTERVAL = timedelta(milliseconds=500)
+_MCP_TASK_MAX_POLL_INTERVAL = timedelta(seconds=5)
+_MCP_TASK_CANCEL_TIMEOUT = timedelta(seconds=5)
+_MCP_TASK_TERMINAL_STATUSES: frozenset[str] = frozenset({"completed", "failed", "cancelled", "input_required"})
+
+# Total send attempts for a Phase 2 request (initial try + one reconnect-and-retry).
+# A single transient disconnect should not abort a long-running task; sustained outages
+# surface as ``_MCPTaskAbandoned`` after the second failure.
+_MCP_RECONNECT_ATTEMPTS = 2
+
+
+class _MCPTaskAbandoned(ToolExecutionException):
+    """Raised when the remote MCP task may still be running and must be cancelled.
+
+    Subclass of ToolExecutionException so callers see a normal tool failure.
+    """
+
+
+class _MCPDeadlineExpired(Exception):
+    """Internal marker for ``max_task_wait`` expiry; distinct from inner TimeoutError."""
+
+
+@experimental(feature_id=ExperimentalFeature.MCP_LONG_RUNNING_TASKS)
+@dataclass(frozen=True)
+class MCPTaskOptions:
+    """Options controlling how MCPTool drives the MCP long-running task lifecycle.
+
+    When an MCP server advertises a tool with ``execution.taskSupport == "required"``,
+    the framework transparently drives the SEP-2663 ``tools/call`` → ``tasks/get``
+    (polled) → ``tasks/result`` lifecycle so the agent sees a normal tool result.
+
+    Instances are immutable; replace the whole object via
+    ``MCPTool.task_options = MCPTaskOptions(...)`` to change behavior.
+
+    Attributes:
+        default_ttl: Optional task-record retention time forwarded to the server as
+            ``params.task.ttl`` (milliseconds, integer). The server keeps the task
+            record around this long after the task reaches a terminal status so the
+            client can still call ``tasks/get`` / ``tasks/result``; it does not
+            cancel a running task. When ``None``, the server applies its own default.
+            Must be positive if set (zero would expire the record before any client
+            could read it).
+        cancel_remote_task_on_local_cancellation: If True (default), a local
+            cancellation of the awaiting coroutine triggers a best-effort
+            ``tasks/cancel`` on the server before re-raising ``CancelledError``.
+            Only gates ``CancelledError``; abandonment paths (max-wait,
+            unrecoverable poll errors, lost connection after task_id is known)
+            always cancel regardless of this flag.
+        max_task_wait: Optional client-side deadline for the whole post-create
+            lifecycle (poll + result fetch). When exceeded, raises
+            ``ToolExecutionException`` and fires a best-effort ``tasks/cancel``.
+            ``None`` (default) means no client-side bound. Must be positive if set.
+    """
+
+    default_ttl: timedelta | None = None
+    cancel_remote_task_on_local_cancellation: bool = True
+    max_task_wait: timedelta | None = None
+
+    def __post_init__(self) -> None:
+        if self.default_ttl is not None and self.default_ttl.total_seconds() <= 0:
+            raise ValueError("MCPTaskOptions.default_ttl must be positive.")
+        if self.max_task_wait is not None and self.max_task_wait.total_seconds() <= 0:
+            raise ValueError("MCPTaskOptions.max_task_wait must be positive.")
+
+
 def streamable_http_client(*args: Any, **kwargs: Any) -> _AsyncGeneratorContextManager[Any, None]:
     """Lazily import the MCP streamable HTTP transport."""
     try:
@@ -217,6 +287,7 @@ class MCPTool:
         request_timeout: int | None = None,
         client: SupportsChatGetResponse | None = None,
         additional_properties: dict[str, Any] | None = None,
+        task_options: MCPTaskOptions | None = None,
     ) -> None:
         """Initialize the MCP Tool base.
 
@@ -248,6 +319,9 @@ class MCPTool:
             request_timeout: Timeout in seconds for MCP requests.
             client: A chat client for sampling callbacks.
             additional_properties: Additional properties for the tool.
+            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.
         """
         self.name = name
         self.description = description or ""
@@ -259,6 +333,10 @@ class MCPTool:
         self.parse_tool_results = parse_tool_results
         self.load_prompts_flag = load_prompts
         self.parse_prompt_results = parse_prompt_results
+        # Defer constructing the default MCPTaskOptions so the experimental warning
+        # only fires when LRO is actually engaged (lazy-resolved by _effective_task_options).
+        self._task_options_explicit: MCPTaskOptions | None = task_options
+        self._task_options_default: MCPTaskOptions | None = None
         self._exit_stack = AsyncExitStack()
         self._lifecycle_lock = asyncio.Lock()
         self._lifecycle_request_lock = asyncio.Lock()
@@ -270,6 +348,7 @@ class MCPTool:
         self.client = client
         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.is_connected: bool = False
         self._tools_loaded: bool = False
         self._prompts_loaded: bool = False
@@ -1131,6 +1210,7 @@ class MCPTool:
         # Track existing function names to prevent duplicates
         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] = {}
 
         params: types.PaginatedRequestParams | None = None
         while True:
@@ -1168,6 +1248,10 @@ class MCPTool:
                 if tool.meta is not None:
                     tool_call_meta_by_name[tool.name] = dict(tool.meta)
 
+                task_support = getattr(getattr(tool, "execution", None), "taskSupport", None)
+                if task_support is not None:
+                    tool_task_support_by_name[tool.name] = task_support
+
                 normalized_name = _normalize_mcp_name(tool.name)
                 local_name = _build_prefixed_mcp_name(normalized_name, self.tool_name_prefix)
 
@@ -1216,6 +1300,7 @@ class MCPTool:
             params = types.PaginatedRequestParams(cursor=tool_list.nextCursor)
 
         self._tool_call_meta_by_name = tool_call_meta_by_name
+        self._tool_task_support_by_name = tool_task_support_by_name
 
     async def _close_on_owner(self) -> None:
         # Cancel any pending reload tasks before tearing down the session.
@@ -1292,6 +1377,29 @@ class MCPTool:
                     inner_exception=ex,
                 ) from ex
 
+    def _effective_task_options(self) -> MCPTaskOptions:
+        """Return the effective MCPTaskOptions, lazily constructing defaults on first use.
+
+        Defers the implicit ``MCPTaskOptions()`` so the experimental warning only
+        fires when LRO is actually engaged (server advertises ``taskSupport=required``).
+        """
+        explicit = self._task_options_explicit
+        if explicit is not None:
+            return explicit
+        if self._task_options_default is None:
+            self._task_options_default = MCPTaskOptions()
+        return self._task_options_default
+
+    @property
+    def task_options(self) -> MCPTaskOptions:
+        """The effective MCPTaskOptions for this tool (lazy defaults)."""
+        return self._effective_task_options()
+
+    @task_options.setter
+    def task_options(self, value: MCPTaskOptions | None) -> None:
+        self._task_options_explicit = value
+        self._task_options_default = None
+
     async def call_tool(self, tool_name: str, **kwargs: Any) -> str | list[Content]:
         """Call a tool with the given arguments.
 
@@ -1322,47 +1430,12 @@ class MCPTool:
                 "Tools are not loaded for this server, please set load_tools=True in the constructor."
             )
 
-        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):
-            raise ToolExecutionException("MCP tool metadata provided via _meta must be a dict.")
-        if isinstance(raw_user_meta, dict):
-            raw_user_meta_dict = cast(Mapping[object, object], raw_user_meta)
-            user_meta = {}
-            for key, value in raw_user_meta_dict.items():
-                if not isinstance(key, str):
-                    raise ToolExecutionException("MCP tool metadata provided via _meta must use string keys.")
-                user_meta[key] = value
+        # Tools advertising taskSupport == "required" cannot complete via plain tools/call;
+        # route through the long-running task lifecycle transparently.
+        if self._tool_task_support_by_name.get(tool_name) == "required":
+            return await self.call_tool_as_task(tool_name, **kwargs)
 
-        # 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
-            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.
-        tool_meta = self._tool_call_meta_by_name.get(tool_name)
-        request_meta = dict(tool_meta) if tool_meta is not None else None
-        if user_meta is not None:
-            request_meta = {**(request_meta or {}), **user_meta}
-        meta = _inject_otel_into_mcp_meta(request_meta)
+        filtered_kwargs, meta = self._prepare_call_kwargs(tool_name, kwargs)
 
         parser = self.parse_tool_results or self._parse_tool_result_from_mcp
         # Try the operation, reconnecting once if the connection is closed
@@ -1411,6 +1484,479 @@ 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 _prepare_call_kwargs(
+        self, tool_name: str, kwargs: dict[str, Any]
+    ) -> tuple[dict[str, Any], dict[str, Any] | None]:
+        """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):
+            raise ToolExecutionException("MCP tool metadata provided via _meta must be a dict.")
+        if isinstance(raw_user_meta, dict):
+            raw_user_meta_dict = cast(Mapping[object, object], raw_user_meta)
+            user_meta = {}
+            for key, value in raw_user_meta_dict.items():
+                if not isinstance(key, str):
+                    raise ToolExecutionException("MCP tool metadata provided via _meta must use string keys.")
+                user_meta[key] = value
+
+        # 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
+            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.
+        tool_meta = self._tool_call_meta_by_name.get(tool_name)
+        request_meta = dict(tool_meta) if tool_meta is not None else None
+        if user_meta is not None:
+            request_meta = {**(request_meta or {}), **user_meta}
+        meta = _inject_otel_into_mcp_meta(request_meta)
+        return filtered_kwargs, meta
+
+    async def call_tool_as_task(self, tool_name: str, **kwargs: Any) -> str | list[Content]:
+        """Call an MCP tool via the long-running task lifecycle (SEP-2663).
+
+        Issues an augmented ``tools/call`` with ``params.task`` set from
+        ``self.task_options``, then polls ``tasks/get`` until the server reports a
+        terminal status. On ``completed`` the payload is fetched via ``tasks/result``,
+        validated as a ``CallToolResult`` and parsed identically to :meth:`call_tool`.
+
+        Local cancellation triggers a best-effort ``tasks/cancel`` (controlled by
+        :attr:`MCPTaskOptions.cancel_remote_task_on_local_cancellation`) before
+        ``asyncio.CancelledError`` is re-raised.
+
+        Args:
+            tool_name: The remote MCP tool name.
+
+        Keyword Args:
+            kwargs: Arguments forwarded to the tool. See :meth:`call_tool` for the
+                framework kwargs that are filtered out.
+
+        Returns:
+            A list of Content items (or a string when a custom ``parse_tool_results``
+            callback is configured).
+        """
+        from anyio import ClosedResourceError
+        from mcp.shared.exceptions import McpError
+
+        if not self.load_tools_flag:
+            raise ToolExecutionException(
+                "Tools are not loaded for this server, please set load_tools=True in the constructor."
+            )
+
+        filtered_kwargs, meta = self._prepare_call_kwargs(tool_name, kwargs)
+        parser = self.parse_tool_results or self._parse_tool_result_from_mcp
+
+        # Submit the task: issue augmented tools/call. Do NOT retry on connection loss here:
+        # the server may have accepted the request and created a task before the
+        # response was lost, so retrying could start the long-running operation twice.
+        # Reconnect-and-retry is only safe after the task_id is known.
+        try:
+            task_id, fallback_result = await self._call_tool_as_task_create(tool_name, filtered_kwargs, meta)
+        except (ClosedResourceError, McpError) as ex:
+            if not self._is_connection_lost(ex):
+                error_message = ex.error.message if isinstance(ex, McpError) else str(ex)
+                raise ToolExecutionException(error_message, inner_exception=ex) from ex
+            raise ToolExecutionException(
+                f"Failed to call tool '{tool_name}' - connection lost; task state unknown.",
+                inner_exception=ex,
+            ) from ex
+        except ToolExecutionException:
+            raise
+        except Exception as ex:
+            raise ToolExecutionException(f"Failed to call tool '{tool_name}'.", inner_exception=ex) from ex
+
+        # Server returned a CallToolResult (no task created) or fell back to plain tools/call.
+        if fallback_result is not None:
+            if fallback_result.isError:
+                parsed = parser(fallback_result)
+                text = (
+                    "\n".join(c.text for c in parsed if c.type == "text" and c.text)
+                    if isinstance(parsed, list)
+                    else str(parsed)
+                )
+                raise ToolExecutionException(text or str(parsed))
+            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}'."
+            )
+
+        # 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.
+        opts = self._effective_task_options()
+        max_wait_s = opts.max_task_wait.total_seconds() if opts.max_task_wait is not None else None
+
+        async def _await_task_completion() -> str | list[Content]:
+            terminal = await self._poll_task_until_terminal(task_id)
+            return await self._handle_terminal_task(tool_name, task_id, terminal, parser)
+
+        try:
+            if max_wait_s is not None:
+                try:
+                    result = await self._await_with_deadline(_await_task_completion(), max_wait_s)
+                    return cast("str | list[Content]", result)
+                except _MCPDeadlineExpired as ex:
+                    self._spawn_best_effort_cancel(task_id)
+                    raise ToolExecutionException(
+                        f"MCP task '{task_id}' exceeded max_task_wait of {max_wait_s}s.",
+                        inner_exception=ex,
+                    ) from ex
+            else:
+                return await _await_task_completion()
+        except asyncio.CancelledError:
+            if opts.cancel_remote_task_on_local_cancellation:
+                self._spawn_best_effort_cancel(task_id)
+            raise
+        except _MCPTaskAbandoned:
+            # Pre-terminal abandonment (hard poll error, malformed get, second
+            # disconnect, reconnect failure): cancel + re-raise as plain
+            # ToolExecutionException to the function-calling loop.
+            self._spawn_best_effort_cancel(task_id)
+            raise
+        # Plain ToolExecutionException from terminal failures (failed/cancelled/
+        # input_required, completed+isError, malformed result post-completion)
+        # propagates without cancel — server is already done.
+
+    async def _call_tool_as_task_create(
+        self, tool_name: str, arguments: dict[str, Any], meta: dict[str, Any] | None
+    ) -> tuple[str | None, types.CallToolResult | None]:
+        """Send the augmented tools/call.
+
+        Returns ``(task_id, None)`` when the server created a task,
+        ``(None, CallToolResult)`` when it returned a non-task result, falling back
+        to plain ``tools/call`` if the server rejects the ``task`` field outright.
+        """
+        from mcp import types
+        from mcp.shared.exceptions import McpError
+        from pydantic import ValidationError
+
+        opts = self._effective_task_options()
+        ttl_ms: int | None = None
+        if opts.default_ttl is not None:
+            ttl_ms = int(opts.default_ttl.total_seconds() * 1000)
+        # Always send TaskMetadata to mark the call as task-augmented; ttl may be omitted.
+        task_metadata = types.TaskMetadata(ttl=ttl_ms)
+
+        request_meta = types.RequestParams.Meta(**meta) if meta else None
+        params = types.CallToolRequestParams(
+            name=tool_name,
+            arguments=arguments,
+            task=task_metadata,
+            _meta=request_meta,  # type: ignore[call-arg]
+        )
+        request = types.ClientRequest(types.CallToolRequest(params=params))
+
+        # Use the lenient Result type so we can extract the task_id even when
+        # the strict CreateTaskResult schema rejects the payload (the MCP Python
+        # SDK requires Task.ttl, but servers may legitimately omit it).
+        try:
+            lenient = await self.session.send_request(  # type: ignore[union-attr]
+                request,
+                types.Result,
+            )
+        except McpError as ex:
+            if ex.error.code not in (types.METHOD_NOT_FOUND, types.INVALID_PARAMS):
+                raise
+            logger.debug(
+                "Server rejected augmented tools/call for '%s' (code=%s); falling back.",
+                tool_name,
+                ex.error.code,
+            )
+            fallback = await self.session.call_tool(tool_name, arguments=arguments, meta=meta)  # type: ignore[union-attr]
+            return None, fallback
+
+        # Inspect the raw payload: a CreateTaskResult carries `task.taskId`;
+        # a legacy CallToolResult carries `content` and/or `isError`.
+        raw: dict[str, Any] = lenient.model_dump(by_alias=True, exclude_none=True)
+        raw.pop("_meta", None)
+
+        task_field = raw.get("task")
+        if isinstance(task_field, dict):
+            task_id_val = cast(dict[str, Any], task_field).get("taskId")
+            if isinstance(task_id_val, str):
+                return task_id_val, None
+
+        try:
+            legacy = types.CallToolResult.model_validate(raw)
+        except ValidationError as ex:
+            # Augmented call succeeded server-side; re-issuing a plain tools/call
+            # could double-execute a side-effecting tool.
+            raise ToolExecutionException(
+                f"MCP server returned an unparseable response to augmented tools/call "
+                f"for '{tool_name}'; cannot safely retry (server may have started the operation).",
+                inner_exception=ex,
+            ) from ex
+
+        return None, legacy
+
+    async def _poll_task_until_terminal(self, task_id: str) -> types.GetTaskResult:
+        """Poll ``tasks/get`` until the task reaches a terminal status."""
+        import httpx
+        from mcp import types
+        from mcp.shared.exceptions import McpError
+
+        # SDK raises McpError(code=httpx.REQUEST_TIMEOUT=408) on session read timeout.
+        transient_codes: frozenset[int] = frozenset({int(httpx.codes.REQUEST_TIMEOUT)})
+
+        while True:
+            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(
+                    request, types.Result, operation="tasks/get", task_id=task_id
+                )
+            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
+                    )
+                    await asyncio.sleep(_MCP_TASK_MIN_POLL_INTERVAL.total_seconds())
+                    continue
+                # Hard server error mid-poll: task may still be running.
+                raise _MCPTaskAbandoned(ex.error.message, inner_exception=ex) from ex
+
+            try:
+                snapshot = self._coerce_get_task_result(lenient, task_id)
+            except ToolExecutionException as ex:
+                # Malformed tasks/get response; task may still be running.
+                raise _MCPTaskAbandoned(str(ex), inner_exception=ex) from ex
+
+            if snapshot.status in _MCP_TASK_TERMINAL_STATUSES:
+                return snapshot
+
+            await asyncio.sleep(self._compute_poll_delay(snapshot.pollInterval).total_seconds())
+
+    @staticmethod
+    def _coerce_get_task_result(lenient: types.Result, task_id: str) -> types.GetTaskResult:
+        """Coerce a lenient Result into GetTaskResult, defaulting ``ttl`` when absent."""
+        from mcp import types
+
+        raw = lenient.model_dump(by_alias=True, exclude_none=True)
+        raw.pop("_meta", None)
+        raw.setdefault("ttl", None)
+        try:
+            return types.GetTaskResult.model_validate(raw)
+        except Exception as ex:
+            raise ToolExecutionException(
+                f"MCP server returned a malformed tasks/get response for task '{task_id}'.",
+                inner_exception=ex,
+            ) from ex
+
+    @staticmethod
+    def _compute_poll_delay(server_interval_ms: int | None) -> timedelta:
+        """Clamp the server-suggested poll interval to ``[min, max]``."""
+        if server_interval_ms is None or server_interval_ms <= 0:
+            return _MCP_TASK_MIN_POLL_INTERVAL
+        suggested = timedelta(milliseconds=server_interval_ms)
+        if suggested < _MCP_TASK_MIN_POLL_INTERVAL:
+            return _MCP_TASK_MIN_POLL_INTERVAL
+        if suggested > _MCP_TASK_MAX_POLL_INTERVAL:
+            return _MCP_TASK_MAX_POLL_INTERVAL
+        return suggested
+
+    async def _handle_terminal_task(
+        self,
+        tool_name: str,
+        task_id: str,
+        snapshot: types.GetTaskResult,
+        parser: Callable[[types.CallToolResult], str | list[Content]],
+    ) -> str | list[Content]:
+        """Map a terminal task snapshot to either a parsed result or an exception."""
+        status = snapshot.status
+        if status == "completed":
+            payload = await self._fetch_task_result(task_id)
+            if payload.isError:
+                parsed = parser(payload)
+                text = (
+                    "\n".join(c.text for c in parsed if c.type == "text" and c.text)
+                    if isinstance(parsed, list)
+                    else str(parsed)
+                )
+                raise ToolExecutionException(text or str(parsed))
+            return parser(payload)
+
+        # Non-completed terminal statuses surface as ToolExecutionException so the
+        # function-calling loop sees a normal failure for tool_name.
+        message = snapshot.statusMessage or f"MCP task ended with status '{status}'."
+        if status == "input_required":
+            # Spec-non-terminal; treated as terminal here because the framework does
+            # not implement the interactive input flow.
+            message = snapshot.statusMessage or "MCP task requires additional input and cannot continue."
+        raise ToolExecutionException(f"Tool '{tool_name}' task {status}: {message}")
+
+    async def _fetch_task_result(self, task_id: str) -> types.CallToolResult:
+        """Send ``tasks/result`` and reinterpret the open-typed payload as a CallToolResult."""
+        from mcp import types
+        from mcp.shared.exceptions import McpError
+        from pydantic import ValidationError
+
+        request = types.ClientRequest(
+            types.GetTaskPayloadRequest(params=types.GetTaskPayloadRequestParams(taskId=task_id))
+        )
+        # Connection-loss retry only via the helper; no transient-code retry — server
+        # has already completed the task, so a slow payload fetch is anomalous.
+        try:
+            payload = await self._send_with_one_reconnect(
+                request, types.GetTaskPayloadResult, operation="tasks/result", task_id=task_id
+            )
+        except McpError as ex:
+            # Server reported completed; a hard fetch error is a plain failure (no cancel).
+            raise ToolExecutionException(ex.error.message, inner_exception=ex) from ex
+
+        # GetTaskPayloadResult carries the tool result via extra fields; reinterpret as CallToolResult.
+        payload_dict = payload.model_dump(by_alias=True, exclude_none=True)
+        payload_dict.pop("_meta", None)
+        try:
+            return types.CallToolResult.model_validate(payload_dict)
+        except ValidationError as ex:
+            # Server reported completed; malformed payload is a plain failure (no cancel needed).
+            raise ToolExecutionException(
+                f"MCP task '{task_id}' result payload could not be parsed as a CallToolResult.",
+                inner_exception=ex,
+            ) from ex
+
+    async def _send_with_one_reconnect(
+        self,
+        request: types.ClientRequest,
+        result_type: type[Any],
+        *,
+        operation: str,
+        task_id: str,
+    ) -> Any:
+        """Send ``request`` with one reconnect-and-retry on connection loss.
+
+        After a second loss (or reconnect failure), raise ``_MCPTaskAbandoned``.
+        Non-connection errors propagate unchanged.
+        """
+        from anyio import ClosedResourceError
+        from mcp.shared.exceptions import McpError
+
+        for attempt in range(_MCP_RECONNECT_ATTEMPTS):
+            try:
+                return await self.session.send_request(request, result_type)  # type: ignore[union-attr]
+            except (ClosedResourceError, McpError) as ex:
+                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
+                    )
+                    try:
+                        await self.connect(reset=True)
+                    except Exception as reconn_ex:
+                        # Reconnect failure: task may still be running.
+                        raise _MCPTaskAbandoned(
+                            "Failed to reconnect to MCP server.", inner_exception=reconn_ex
+                        ) from reconn_ex
+                    continue
+                # Final attempt also lost the connection: task may still be running.
+                raise _MCPTaskAbandoned(
+                    f"MCP connection lost; task state unknown (task_id={task_id}).",
+                    inner_exception=ex,
+                ) from ex
+        raise AssertionError(f"unreachable: {operation} for {task_id}")  # pragma: no cover
+
+    @staticmethod
+    async def _await_with_deadline(coro: Coroutine[Any, Any, Any], timeout_s: float) -> Any:
+        """Await ``coro`` with a deadline; raise ``_MCPDeadlineExpired`` only on deadline.
+
+        Unlike ``asyncio.wait_for``, an ``asyncio.TimeoutError`` raised by ``coro``
+        itself propagates unchanged so callers can distinguish their own deadline
+        from a stray inner timeout.
+        """
+        inner = asyncio.ensure_future(coro)
+        try:
+            done, _pending = await asyncio.wait({inner}, timeout=timeout_s)
+        except BaseException:
+            # Outer caller cancelled (or another exception): cancel inner + drain.
+            inner.cancel()
+            with contextlib.suppress(BaseException):
+                await inner
+            raise
+        if inner in done:
+            return inner.result()
+        # Deadline fired before inner finished.
+        inner.cancel()
+        with contextlib.suppress(BaseException):
+            await inner
+        raise _MCPDeadlineExpired
+
+    def _spawn_best_effort_cancel(self, task_id: str) -> None:
+        """Fire-and-forget ``tasks/cancel`` so local cancellation propagates server-side."""
+        try:
+            loop = asyncio.get_running_loop()
+        except RuntimeError:
+            return
+        cancel_task = loop.create_task(self._try_cancel_task(task_id))
+        # Reuse pending-reload bookkeeping so close-on-owner waits/cancels these too.
+        self._pending_reload_tasks.add(cancel_task)
+        cancel_task.add_done_callback(self._pending_reload_tasks.discard)
+
+    async def _try_cancel_task(self, task_id: str) -> None:
+        """Send ``tasks/cancel``; bounded by ``_MCP_TASK_CANCEL_TIMEOUT``.
+
+        Failures log at warning so unattributed orphan tasks are debuggable.
+        """
+        from mcp import types
+
+        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]
+                timeout=_MCP_TASK_CANCEL_TIMEOUT.total_seconds(),
+            )
+        except asyncio.CancelledError:
+            raise
+        except asyncio.TimeoutError:
+            logger.warning(
+                "Best-effort tasks/cancel for '%s' timed out after %.1fs; "
+                "remote task may still be running.",
+                task_id,
+                _MCP_TASK_CANCEL_TIMEOUT.total_seconds(),
+            )
+        except Exception:
+            logger.warning(
+                "Best-effort tasks/cancel for '%s' failed; remote task may still be running.",
+                task_id,
+                exc_info=True,
+            )
+
+    @staticmethod
+    def _is_connection_lost(ex: BaseException) -> bool:
+        """Return True if *ex* indicates the MCP transport was torn down."""
+        from anyio import ClosedResourceError
+        from mcp.shared.exceptions import McpError
+
+        if isinstance(ex, ClosedResourceError):
+            return True
+        if isinstance(ex, McpError):
+            return "session terminated" in ex.error.message.lower()
+        return False
+
     async def get_prompt(self, prompt_name: str, **kwargs: Any) -> str:
         """Call a prompt with the given arguments.
 
@@ -1554,6 +2100,7 @@ class MCPStdioTool(MCPTool):
         encoding: str | None = None,
         client: SupportsChatGetResponse | None = None,
         additional_properties: dict[str, Any] | None = None,
+        task_options: MCPTaskOptions | None = None,
         **kwargs: Any,
     ) -> None:
         """Initialize the MCP stdio tool.
@@ -1598,6 +2145,8 @@ class MCPStdioTool(MCPTool):
             env: The environment variables to set for the command.
             encoding: The encoding to use for the command output.
             client: The chat client to use for sampling.
+            task_options: Options for tools that advertise
+                ``execution.taskSupport == "required"``. See :class:`MCPTaskOptions`.
             kwargs: Any extra arguments to pass to the stdio client.
         """
         super().__init__(
@@ -1614,6 +2163,7 @@ class MCPStdioTool(MCPTool):
             load_prompts=load_prompts,
             parse_prompt_results=parse_prompt_results,
             request_timeout=request_timeout,
+            task_options=task_options,
         )
         self.command = command
         self.args = args or []
@@ -1687,6 +2237,7 @@ class MCPStreamableHTTPTool(MCPTool):
         additional_properties: dict[str, Any] | None = None,
         http_client: AsyncClient | None = None,
         header_provider: Callable[[dict[str, Any]], dict[str, str]] | None = None,
+        task_options: MCPTaskOptions | None = None,
         **kwargs: Any,
     ) -> None:
         """Initialize the MCP streamable HTTP tool.
@@ -1739,6 +2290,8 @@ class MCPStreamableHTTPTool(MCPTool):
                 of HTTP headers to inject into every outbound request to the MCP server.
                 Use this to forward per-request context (e.g. authentication tokens set in
                 agent middleware) without creating a separate ``httpx.AsyncClient``.
+            task_options: Options for tools that advertise
+                ``execution.taskSupport == "required"``. See :class:`MCPTaskOptions`.
             kwargs: Additional keyword arguments (accepted for backward compatibility but not used).
         """
         super().__init__(
@@ -1755,6 +2308,7 @@ class MCPStreamableHTTPTool(MCPTool):
             load_prompts=load_prompts,
             parse_prompt_results=parse_prompt_results,
             request_timeout=request_timeout,
+            task_options=task_options,
         )
         self.url = url
         self.terminate_on_close = terminate_on_close
@@ -1862,6 +2416,7 @@ class MCPWebsocketTool(MCPTool):
         allowed_tools: Collection[str] | None = None,
         client: SupportsChatGetResponse | None = None,
         additional_properties: dict[str, Any] | None = None,
+        task_options: MCPTaskOptions | None = None,
         **kwargs: Any,
     ) -> None:
         """Initialize the MCP WebSocket tool.
@@ -1904,6 +2459,8 @@ class MCPWebsocketTool(MCPTool):
             allowed_tools: A list of tools that are allowed to use this tool.
             additional_properties: Additional properties.
             client: The chat client to use for sampling.
+            task_options: Options for tools that advertise
+                ``execution.taskSupport == "required"``. See :class:`MCPTaskOptions`.
             kwargs: Any extra arguments to pass to the WebSocket client.
         """
         super().__init__(
@@ -1920,6 +2477,7 @@ class MCPWebsocketTool(MCPTool):
             load_prompts=load_prompts,
             parse_prompt_results=parse_prompt_results,
             request_timeout=request_timeout,
+            task_options=task_options,
         )
         self.url = url
         self._client_kwargs = kwargs

python/packages/core/agent_framework/_tools.py

@@ -292,7 +292,6 @@ class FunctionTool(SerializationMixin):
         "_cached_parameters",
         "_input_schema",
         "_schema_supplied",
-        "_invoke_sync_on_event_loop",
     }
 
     def __init__(
@@ -367,7 +366,6 @@ class FunctionTool(SerializationMixin):
         self.description = description
         self.kind = kind
         self.additional_properties = additional_properties
-        self._invoke_sync_on_event_loop = False
         for key, value in kwargs.items():
             setattr(self, key, value)
 
@@ -539,16 +537,6 @@ class FunctionTool(SerializationMixin):
             self.invocation_exception_count += 1
             raise
 
-    async def _invoke_function(self, call_kwargs: Mapping[str, Any]) -> Any:
-        """Run sync tools off the event loop during async invocation."""
-        func = self.func.func if isinstance(self.func, FunctionTool) else self.func
-        if inspect.iscoroutinefunction(func) or getattr(self, "_invoke_sync_on_event_loop", False):
-            res = self.__call__(**call_kwargs)
-            return await res if inspect.isawaitable(res) else res
-
-        res = await asyncio.to_thread(self.__call__, **call_kwargs)
-        return await res if inspect.isawaitable(res) else res
-
     @overload
     async def invoke(
         self,
@@ -691,7 +679,8 @@ class FunctionTool(SerializationMixin):
         if not OBSERVABILITY_SETTINGS.ENABLED:  # type: ignore[name-defined]
             logger.info(f"Function name: {self.name}")
             logger.debug(f"Function arguments: {observable_kwargs}")
-            result = await self._invoke_function(call_kwargs)
+            res = self.__call__(**call_kwargs)
+            result = await res if inspect.isawaitable(res) else res
             if skip_parsing:
                 logger.info(f"Function {self.name} succeeded.")
                 logger.debug(f"Function result: {type(result).__name__}")
@@ -741,7 +730,8 @@ class FunctionTool(SerializationMixin):
             start_time_stamp = perf_counter()
             end_time_stamp: float | None = None
             try:
-                result = await self._invoke_function(call_kwargs)
+                res = self.__call__(**call_kwargs)
+                result = await res if inspect.isawaitable(res) else res
                 end_time_stamp = perf_counter()
             except Exception as exception:
                 end_time_stamp = perf_counter()

python/packages/core/pyproject.toml

@@ -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 = "1.8.0"
+version = "1.7.0"
 license-files = ["LICENSE"]
 urls.homepage = "https://aka.ms/agent-framework"
 urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"

python/packages/core/tests/core/test_clients.py

@@ -11,14 +11,10 @@ from agent_framework import (
     GROUP_TOKEN_COUNT_KEY,
     BaseChatClient,
     ChatResponse,
-    ChatResponseUpdate,
-    Content,
     Message,
     SlidingWindowStrategy,
     SupportsChatGetResponse,
-    ToolResultCompactionStrategy,
     TruncationStrategy,
-    tool,
 )
 
 
@@ -262,196 +258,6 @@ async def test_base_client_default_tokenizer_without_strategy_annotates_messages
     assert captured_token_counts == [[19, 19]]
 
 
-def _tool_call_response(call_id: str, location: str) -> ChatResponse:
-    return ChatResponse(
-        messages=Message(
-            role="assistant",
-            contents=[
-                Content.from_function_call(
-                    call_id=call_id,
-                    name="lookup_weather",
-                    arguments=f'{{"location": "{location}"}}',
-                )
-            ],
-        ),
-        response_id=f"resp_{call_id}",
-    )
-
-
-def _is_tool_result_summary(message: Message) -> bool:
-    text = message.text or ""
-    return message.role == "assistant" and text.startswith("[Tool results:")
-
-
-async def test_function_loop_persists_inserted_summaries_across_iterations(
-    chat_client_base: SupportsChatGetResponse,
-) -> None:
-    # Regression test for #4991: compaction inserts summary messages and excludes the
-    # originals. Across tool-loop iterations the exclusion flags persisted (shared Message
-    # objects) but the inserted summaries were dropped (they only lived on a throwaway copy),
-    # so older tool groups were silently lost with no summary representing them.
-    chat_client_base.function_invocation_configuration["enabled"] = True  # type: ignore[attr-defined]
-    chat_client_base.function_invocation_configuration["max_iterations"] = 3  # type: ignore[attr-defined]
-    chat_client_base.compaction_strategy = ToolResultCompactionStrategy(keep_last_tool_call_groups=1)  # type: ignore[attr-defined]
-
-    @tool(name="lookup_weather", approval_mode="never_require")
-    def lookup_weather(location: str) -> str:
-        return f"Weather in {location}: sunny"
-
-    chat_client_base.run_responses = [  # type: ignore[attr-defined]
-        _tool_call_response("call_1", "London"),
-        _tool_call_response("call_2", "Paris"),
-        _tool_call_response("call_3", "Tokyo"),
-    ]
-
-    captured_inputs: list[list[Message]] = []
-    original = chat_client_base._get_non_streaming_response  # type: ignore[attr-defined]
-
-    async def _capture(
-        *,
-        messages: list[Message],
-        options: dict[str, Any],
-        **kwargs: Any,
-    ) -> ChatResponse:
-        captured_inputs.append(list(messages))
-        return await original(messages=messages, options=options, **kwargs)
-
-    chat_client_base._get_non_streaming_response = _capture  # type: ignore[attr-defined,method-assign]
-
-    await chat_client_base.get_response(
-        [Message(role="user", contents=["What is the weather in London?"])],
-        options={"tools": [lookup_weather]},  # type: ignore[typeddict-unknown-key]
-    )
-
-    # The final model call should represent every compacted tool group with a summary.
-    # Two older tool groups get collapsed (London, Paris) while the last (Tokyo) is kept.
-    final_input = captured_inputs[-1]
-    summaries = [message for message in final_input if _is_tool_result_summary(message)]
-    summary_text = " ".join(message.text or "" for message in summaries)
-
-    assert len(summaries) == 2, [message.text for message in final_input]
-    assert "London" in summary_text
-    assert "Paris" in summary_text
-
-
-def _tool_call_update(call_id: str, location: str) -> list[ChatResponseUpdate]:
-    return [
-        ChatResponseUpdate(
-            contents=[
-                Content.from_function_call(
-                    call_id=call_id,
-                    name="lookup_weather",
-                    arguments=f'{{"location": "{location}"}}',
-                )
-            ],
-            role="assistant",
-            finish_reason="stop",
-            response_id=f"resp_{call_id}",
-        )
-    ]
-
-
-async def test_function_loop_persists_inserted_summaries_across_iterations_streaming(
-    chat_client_base: SupportsChatGetResponse,
-) -> None:
-    # Streaming counterpart of the #4991 regression test: the summary persistence fix in
-    # ``_prepare_messages_for_model_call`` must cover the streaming tool loop too.
-    chat_client_base.function_invocation_configuration["enabled"] = True  # type: ignore[attr-defined]
-    chat_client_base.function_invocation_configuration["max_iterations"] = 3  # type: ignore[attr-defined]
-    chat_client_base.compaction_strategy = ToolResultCompactionStrategy(keep_last_tool_call_groups=1)  # type: ignore[attr-defined]
-
-    @tool(name="lookup_weather", approval_mode="never_require")
-    def lookup_weather(location: str) -> str:
-        return f"Weather in {location}: sunny"
-
-    chat_client_base.streaming_responses = [  # type: ignore[attr-defined]
-        _tool_call_update("call_1", "London"),
-        _tool_call_update("call_2", "Paris"),
-        _tool_call_update("call_3", "Tokyo"),
-    ]
-
-    captured_inputs: list[list[Message]] = []
-    original = chat_client_base._get_streaming_response  # type: ignore[attr-defined]
-
-    def _capture(
-        *,
-        messages: list[Message],
-        options: dict[str, Any],
-        **kwargs: Any,
-    ):
-        captured_inputs.append(list(messages))
-        return original(messages=messages, options=options, **kwargs)
-
-    chat_client_base._get_streaming_response = _capture  # type: ignore[attr-defined,method-assign]
-
-    stream = chat_client_base.get_response(
-        [Message(role="user", contents=["What is the weather in London?"])],
-        stream=True,
-        options={"tools": [lookup_weather]},  # type: ignore[typeddict-unknown-key]
-    )
-    async for _ in stream:
-        pass
-
-    final_input = captured_inputs[-1]
-    summaries = [message for message in final_input if _is_tool_result_summary(message)]
-    summary_text = " ".join(message.text or "" for message in summaries)
-
-    assert len(summaries) == 2, [message.text for message in final_input]
-    assert "London" in summary_text
-    assert "Paris" in summary_text
-
-
-async def test_function_loop_compaction_conversation_id_mode_does_not_resend_history(
-    chat_client_base: SupportsChatGetResponse,
-) -> None:
-    # In conversation-id mode the server owns prior context, so the tool loop clears
-    # ``prepped_messages`` and only sends the latest message. Compaction must not fight that
-    # by re-inserting summaries or re-sending earlier turns.
-    chat_client_base.function_invocation_configuration["enabled"] = True  # type: ignore[attr-defined]
-    chat_client_base.function_invocation_configuration["max_iterations"] = 3  # type: ignore[attr-defined]
-    chat_client_base.compaction_strategy = ToolResultCompactionStrategy(keep_last_tool_call_groups=1)  # type: ignore[attr-defined]
-
-    @tool(name="lookup_weather", approval_mode="never_require")
-    def lookup_weather(location: str) -> str:
-        return f"Weather in {location}: sunny"
-
-    def _conversation_tool_call(call_id: str, location: str) -> ChatResponse:
-        response = _tool_call_response(call_id, location)
-        response.conversation_id = "conv_1"
-        return response
-
-    chat_client_base.run_responses = [  # type: ignore[attr-defined]
-        _conversation_tool_call("call_1", "London"),
-        _conversation_tool_call("call_2", "Paris"),
-        _conversation_tool_call("call_3", "Tokyo"),
-    ]
-
-    captured_inputs: list[list[Message]] = []
-    original = chat_client_base._get_non_streaming_response  # type: ignore[attr-defined]
-
-    async def _capture(
-        *,
-        messages: list[Message],
-        options: dict[str, Any],
-        **kwargs: Any,
-    ) -> ChatResponse:
-        captured_inputs.append(list(messages))
-        return await original(messages=messages, options=options, **kwargs)
-
-    chat_client_base._get_non_streaming_response = _capture  # type: ignore[attr-defined,method-assign]
-
-    await chat_client_base.get_response(
-        [Message(role="user", contents=["What is the weather in London?"])],
-        options={"tools": [lookup_weather]},  # type: ignore[typeddict-unknown-key]
-    )
-
-    # After the conversation id is established the loop only forwards the latest message,
-    # so subsequent model calls never receive the full history or summary messages.
-    for sent in captured_inputs[1:]:
-        assert len(sent) <= 1, [message.text for message in sent]
-        assert not any(_is_tool_result_summary(message) for message in sent)
-
-
 def test_base_client_as_agent_does_not_copy_client_compaction_defaults(
     chat_client_base: SupportsChatGetResponse,
 ) -> None:

python/packages/core/tests/core/test_compaction.py

@@ -196,64 +196,6 @@ def test_append_compaction_message_annotates_new_message() -> None:
     assert isinstance(_group_id(messages[1]), str)
 
 
-def test_incremental_annotation_assigns_unique_message_ids() -> None:
-    # Regression test for #5237: ``_ensure_message_ids`` assigned ``msg_{index}``
-    # using the position within the slice handed to ``group_messages``. Successive
-    # incremental annotations restart the index at 0, so distinct messages collided
-    # on the same ``message_id``.
-    messages: list[Message] = []
-    for turn in range(4):
-        messages.append(Message(role="user", contents=[f"user {turn}"]))
-        annotate_message_groups(messages)
-        messages.append(Message(role="assistant", contents=[f"assistant {turn}"]))
-        annotate_message_groups(messages)
-
-    message_ids = [message.message_id for message in messages]
-    assert all(message_ids), "every message should receive an id"
-    assert len(set(message_ids)) == len(message_ids), f"duplicate message ids: {message_ids}"
-
-
-def test_ensure_message_ids_avoids_existing_id_collisions() -> None:
-    # An auto-generated ``msg_{index}`` must not collide with an id already present
-    # on another message (user-supplied or assigned by an earlier annotation pass).
-    messages = [
-        Message(role="user", contents=["zero"]),
-        Message(role="assistant", contents=["one"], message_id="msg_2"),
-        Message(role="user", contents=["two"]),
-    ]
-    annotate_message_groups(messages)
-
-    message_ids = [message.message_id for message in messages]
-    assert message_ids[1] == "msg_2"
-    assert len(set(message_ids)) == len(message_ids), f"duplicate message ids: {message_ids}"
-
-
-def test_incremental_annotation_avoids_prefix_id_collision() -> None:
-    # Regression for the PR review on #5237: when only a suffix is re-annotated,
-    # an auto-assigned ``msg_{index}`` in the suffix must not collide with a
-    # preexisting id carried by a message in the *preserved prefix* (a group
-    # before the one re-annotation pulls back to). Otherwise ``_group_id_for``
-    # derives the same group id and merges groups across the boundary.
-    messages = [
-        # Out-of-position, user-supplied id that matches the ``msg_{index}`` the
-        # suffix pass would assign to the appended message below. This message is
-        # two groups back, so it stays outside the re-annotated slice.
-        Message(role="user", contents=["zero"], message_id="msg_2"),
-        Message(role="user", contents=["one"]),
-    ]
-    annotate_message_groups(messages)
-    assert messages[0].message_id == "msg_2"
-    assert messages[1].message_id == "msg_1"
-
-    messages.append(Message(role="user", contents=["two"]))
-    annotate_message_groups(messages, from_index=2)
-
-    message_ids = [message.message_id for message in messages]
-    assert all(message_ids), "every message should receive an id"
-    assert len(set(message_ids)) == len(message_ids), f"duplicate message ids: {message_ids}"
-    assert messages[0].message_id == "msg_2"
-
-
 async def test_truncation_strategy_keeps_system_anchor() -> None:
     messages = [
         Message(role="system", contents=["you are helpful"]),
@@ -542,44 +484,6 @@ async def test_tool_result_compaction_collapses_old_groups_into_summary() -> Non
     assert any(m.role == "tool" for m in projected)
 
 
-async def test_tool_result_compaction_is_idempotent_after_summary_insertion() -> None:
-    """Re-running compaction after a mid-list summary insertion must not duplicate it.
-
-    Mirrors a subsequent tool-loop iteration (issue #4991): the inserted summary and the
-    excluded originals now persist on the same list, so a second annotate + compaction pass
-    over the same groups should be a no-op rather than collapsing the group again.
-    """
-    messages = [
-        Message(role="user", contents=["u"]),
-        _assistant_function_call("call-1"),
-        _tool_result("call-1", "r1"),
-        _assistant_function_call("call-2"),
-        _tool_result("call-2", "r2"),
-        Message(role="assistant", contents=["done"]),
-    ]
-    strategy = ToolResultCompactionStrategy(keep_last_tool_call_groups=1)
-    annotate_message_groups(messages)
-    assert await strategy(messages) is True
-
-    summaries_after_first = [m for m in messages if (m.text or "").startswith("[Tool results:")]
-    assert len(summaries_after_first) == 1
-    summary = summaries_after_first[0]
-    summary_group_ids = _group_unknown_value(summary, SUMMARY_OF_GROUP_IDS_KEY)
-
-    # Second pass over the same (now partially compacted) list.
-    annotate_message_groups(messages)
-    changed = await strategy(messages)
-
-    assert changed is False
-    summaries_after_second = [m for m in messages if (m.text or "").startswith("[Tool results:")]
-    assert len(summaries_after_second) == 1
-    assert _group_unknown_value(summaries_after_second[0], SUMMARY_OF_GROUP_IDS_KEY) == summary_group_ids
-
-    # The kept tool-call group stays atomic and included.
-    projected = included_messages(messages)
-    assert any(m.role == "tool" for m in projected)
-
-
 async def test_tool_result_compaction_zero_collapses_all() -> None:
     """With keep=0, all tool-call groups are collapsed into summaries."""
     messages = [

python/packages/core/tests/core/test_tools.py

@@ -1,6 +1,4 @@
 # Copyright (c) Microsoft. All rights reserved.
-import asyncio
-import threading
 from typing import Annotated, Any, Literal, get_args, get_origin
 from unittest.mock import Mock
 
@@ -1348,45 +1346,6 @@ async def test_invoke_skip_parsing_awaits_async_functions() -> None:
     assert raw == 42
 
 
-async def test_invoke_sync_tool_does_not_block_event_loop() -> None:
-    release_tool = threading.Event()
-    tool_thread_ids: list[int] = []
-    event_loop_thread_id = threading.get_ident()
-
-    @tool
-    def wait_for_release() -> str:
-        tool_thread_ids.append(threading.get_ident())
-        return "released" if release_tool.wait(timeout=0.2) else "timed out"
-
-    async def release_soon() -> None:
-        await asyncio.sleep(0.01)
-        release_tool.set()
-
-    tool_task = asyncio.create_task(wait_for_release.invoke(skip_parsing=True))
-    release_task = asyncio.create_task(release_soon())
-
-    assert await asyncio.wait_for(tool_task, timeout=1) == "released"
-    await release_task
-    assert tool_thread_ids
-    assert tool_thread_ids[0] != event_loop_thread_id
-
-
-async def test_invoke_sync_tool_can_stay_on_event_loop() -> None:
-    event_loop_thread_id = threading.get_ident()
-    tool_thread_ids: list[int] = []
-
-    @tool
-    def needs_event_loop() -> str:
-        tool_thread_ids.append(threading.get_ident())
-        asyncio.get_running_loop()
-        return "ok"
-
-    needs_event_loop._invoke_sync_on_event_loop = True
-
-    assert await needs_event_loop.invoke(skip_parsing=True) == "ok"
-    assert tool_thread_ids == [event_loop_thread_id]
-
-
 async def test_invoke_skip_parsing_bypasses_configured_result_parser() -> None:
     """The tool's own result_parser is bypassed when skip_parsing=True is requested."""
     parser_calls: list[Any] = []

python/packages/foundry/agent_framework_foundry/_agent.py

@@ -191,7 +191,6 @@ class RawFoundryAgentChatClient(  # type: ignore[misc]
         compaction_strategy: CompactionStrategy | None = None,
         tokenizer: TokenizerProtocol | None = None,
         additional_properties: dict[str, Any] | None = None,
-        timeout: float | None = None,
     ) -> None:
         """Initialize a raw Foundry Agent client.
 
@@ -212,8 +211,6 @@ class RawFoundryAgentChatClient(  # type: ignore[misc]
             compaction_strategy: Optional per-client compaction override.
             tokenizer: Optional tokenizer for compaction strategies.
             additional_properties: Additional properties stored on the client instance.
-            timeout: HTTP timeout in seconds for requests. When not provided, the
-                OpenAI SDK default is used (connect: 5s, total: 600s).
         """
         settings = load_settings(
             FoundryAgentSettings,
@@ -263,11 +260,8 @@ class RawFoundryAgentChatClient(  # type: ignore[misc]
             openai_client_kwargs["default_headers"] = dict(default_headers)
         if allow_preview:
             openai_client_kwargs["agent_name"] = self.agent_name
-        openai_client = self.project_client.get_openai_client(**openai_client_kwargs)
-        if timeout is not None:
-            openai_client = openai_client.with_options(timeout=timeout)
         super().__init__(
-            async_client=openai_client,
+            async_client=self.project_client.get_openai_client(**openai_client_kwargs),
             default_headers=default_headers,
             instruction_role=instruction_role,
             compaction_strategy=compaction_strategy,
@@ -543,7 +537,6 @@ class _FoundryAgentChatClient(  # type: ignore[misc]
         additional_properties: dict[str, Any] | None = None,
         middleware: (Sequence[ChatAndFunctionMiddlewareTypes] | None) = None,
         function_invocation_configuration: FunctionInvocationConfiguration | None = None,
-        timeout: float | None = None,
     ) -> None:
         """Initialize a Foundry Agent client with full middleware support.
 
@@ -563,8 +556,6 @@ class _FoundryAgentChatClient(  # type: ignore[misc]
             additional_properties: Additional properties stored on the client instance.
             middleware: Optional sequence of middleware.
             function_invocation_configuration: Optional function invocation configuration.
-            timeout: HTTP timeout in seconds for requests. When not provided, the
-                OpenAI SDK default is used (connect: 5s, total: 600s).
         """
         super().__init__(
             project_endpoint=project_endpoint,
@@ -582,7 +573,6 @@ class _FoundryAgentChatClient(  # type: ignore[misc]
             additional_properties=additional_properties,
             middleware=middleware,
             function_invocation_configuration=function_invocation_configuration,
-            timeout=timeout,
         )
 
 
@@ -635,7 +625,6 @@ class RawFoundryAgent(  # type: ignore[misc]
         compaction_strategy: CompactionStrategy | None = None,
         tokenizer: TokenizerProtocol | None = None,
         additional_properties: Mapping[str, Any] | None = None,
-        timeout: float | None = None,
     ) -> None:
         """Initialize a Foundry Agent.
 
@@ -668,8 +657,6 @@ class RawFoundryAgent(  # type: ignore[misc]
             compaction_strategy: Optional agent-level in-run compaction override.
             tokenizer: Optional agent-level tokenizer override.
             additional_properties: Additional properties stored on the local agent wrapper.
-            timeout: HTTP timeout in seconds for requests. When not provided, the
-                OpenAI SDK default is used (connect: 5s, total: 600s).
         """
         # Create the client
         actual_client_type = client_type or _FoundryAgentChatClient
@@ -688,7 +675,6 @@ class RawFoundryAgent(  # type: ignore[misc]
             "default_headers": default_headers,
             "env_file_path": env_file_path,
             "env_file_encoding": env_file_encoding,
-            "timeout": timeout,
         }
         if function_invocation_configuration is not None:
             if not issubclass(actual_client_type, FunctionInvocationLayer):
@@ -926,7 +912,6 @@ class FoundryAgent(  # type: ignore[misc]
         compaction_strategy: CompactionStrategy | None = None,
         tokenizer: TokenizerProtocol | None = None,
         additional_properties: Mapping[str, Any] | None = None,
-        timeout: float | None = None,
     ) -> None:
         """Initialize a Foundry Agent with full middleware and telemetry.
 
@@ -973,8 +958,6 @@ class FoundryAgent(  # type: ignore[misc]
             compaction_strategy: Optional agent-level in-run compaction override.
             tokenizer: Optional agent-level tokenizer override.
             additional_properties: Additional properties stored on the local agent wrapper.
-            timeout: HTTP timeout in seconds for requests. When not provided, the
-                OpenAI SDK default is used (connect: 5s, total: 600s).
         """
         super().__init__(
             project_endpoint=project_endpoint,
@@ -1000,5 +983,4 @@ class FoundryAgent(  # type: ignore[misc]
             compaction_strategy=compaction_strategy,
             tokenizer=tokenizer,
             additional_properties=additional_properties,
-            timeout=timeout,
         )

python/packages/foundry/pyproject.toml

@@ -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 = "1.8.0"
+version = "1.7.0"
 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>=1.8.0,<2",
-    "agent-framework-openai>=1.8.0,<2",
+    "agent-framework-core>=1.7.0,<2",
+    "agent-framework-openai>=1.7.0,<2",
     "azure-ai-inference>=1.0.0b9,<1.0.0b10",
     "azure-ai-projects>=2.1.0,<3.0",
 ]

python/packages/foundry/tests/foundry/test_foundry_agent.py

@@ -109,67 +109,9 @@ def test_raw_foundry_agent_chat_client_init_uses_explicit_parameters() -> None:
     assert "compaction_strategy" in signature.parameters
     assert "tokenizer" in signature.parameters
     assert "additional_properties" in signature.parameters
-    assert "timeout" in signature.parameters
     assert all(parameter.kind != inspect.Parameter.VAR_KEYWORD for parameter in signature.parameters.values())
 
 
-def test_raw_foundry_agent_chat_client_init_applies_timeout_to_openai_client() -> None:
-    """Test that timeout is applied via with_options without mutating the shared OpenAI client."""
-
-    mock_project = MagicMock()
-    openai_client_mock = MagicMock()
-    openai_client_mock.timeout = 5.0
-    mock_project.get_openai_client.return_value = openai_client_mock
-
-    client = RawFoundryAgentChatClient(
-        project_client=mock_project,
-        agent_name="test-agent",
-        timeout=60.0,
-    )
-
-    openai_client_mock.with_options.assert_called_once_with(timeout=60.0)
-    assert openai_client_mock.timeout == 5.0, "Original shared client must not be mutated"
-    assert client.client is openai_client_mock.with_options.return_value
-
-
-def test_raw_foundry_agent_chat_client_init_timeout_none_leaves_client_unchanged() -> None:
-    """Test that timeout=None does not call with_options and leaves the shared client intact."""
-
-    mock_project = MagicMock()
-    openai_client_mock = MagicMock()
-    openai_client_mock.timeout = 5.0
-    mock_project.get_openai_client.return_value = openai_client_mock
-
-    RawFoundryAgentChatClient(
-        project_client=mock_project,
-        agent_name="test-agent",
-        timeout=None,
-    )
-
-    openai_client_mock.with_options.assert_not_called()
-    assert openai_client_mock.timeout == 5.0
-
-
-def test_raw_foundry_agent_chat_client_init_applies_timeout_with_preview_enabled() -> None:
-    """Test that timeout uses with_options even when allow_preview=True (hosted agent path)."""
-
-    mock_project = MagicMock()
-    openai_client_mock = MagicMock()
-    openai_client_mock.timeout = 5.0
-    mock_project.get_openai_client.return_value = openai_client_mock
-
-    client = RawFoundryAgentChatClient(
-        project_client=mock_project,
-        agent_name="hosted-agent",
-        allow_preview=True,
-        timeout=120.0,
-    )
-
-    openai_client_mock.with_options.assert_called_once_with(timeout=120.0)
-    assert openai_client_mock.timeout == 5.0, "Original shared client must not be mutated"
-    assert client.client is openai_client_mock.with_options.return_value
-
-
 def test_raw_foundry_agent_chat_client_as_agent_preserves_client_type() -> None:
     """Test that as_agent() wraps the client in FoundryAgent using the same client class."""
 
@@ -610,29 +552,9 @@ def test_foundry_agent_chat_client_init_uses_explicit_parameters() -> None:
     assert "compaction_strategy" in signature.parameters
     assert "tokenizer" in signature.parameters
     assert "additional_properties" in signature.parameters
-    assert "timeout" in signature.parameters
     assert all(parameter.kind != inspect.Parameter.VAR_KEYWORD for parameter in signature.parameters.values())
 
 
-def test_foundry_agent_chat_client_init_propagates_timeout() -> None:
-    """Test that _FoundryAgentChatClient calls with_options instead of mutating the shared client."""
-
-    mock_project = MagicMock()
-    openai_client_mock = MagicMock()
-    openai_client_mock.timeout = 5.0
-    mock_project.get_openai_client.return_value = openai_client_mock
-
-    client = _FoundryAgentChatClient(
-        project_client=mock_project,
-        agent_name="test-agent",
-        timeout=45.0,
-    )
-
-    openai_client_mock.with_options.assert_called_once_with(timeout=45.0)
-    assert openai_client_mock.timeout == 5.0, "Original shared client must not be mutated"
-    assert client.client is openai_client_mock.with_options.return_value
-
-
 def test_raw_foundry_agent_init_creates_client() -> None:
     """Test that RawFoundryAgent creates a client internally."""
 
@@ -707,7 +629,6 @@ def test_raw_foundry_agent_init_uses_explicit_parameters() -> None:
     assert "compaction_strategy" in signature.parameters
     assert "tokenizer" in signature.parameters
     assert "additional_properties" in signature.parameters
-    assert "timeout" in signature.parameters
     assert all(parameter.kind != inspect.Parameter.VAR_KEYWORD for parameter in signature.parameters.values())
 
 
@@ -720,47 +641,9 @@ def test_foundry_agent_init_uses_explicit_parameters() -> None:
     assert "compaction_strategy" in signature.parameters
     assert "tokenizer" in signature.parameters
     assert "additional_properties" in signature.parameters
-    assert "timeout" in signature.parameters
     assert all(parameter.kind != inspect.Parameter.VAR_KEYWORD for parameter in signature.parameters.values())
 
 
-def test_foundry_agent_init_propagates_timeout_to_openai_client() -> None:
-    """Test that FoundryAgent uses with_options instead of mutating the shared OpenAI client."""
-
-    mock_project = MagicMock()
-    openai_client_mock = MagicMock()
-    openai_client_mock.timeout = 5.0
-    mock_project.get_openai_client.return_value = openai_client_mock
-
-    agent = FoundryAgent(
-        project_client=mock_project,
-        agent_name="test-agent",
-        timeout=90.0,
-    )
-
-    openai_client_mock.with_options.assert_called_once_with(timeout=90.0)
-    assert openai_client_mock.timeout == 5.0, "Original shared client must not be mutated"
-    assert agent.client.client is openai_client_mock.with_options.return_value
-
-
-def test_foundry_agent_init_timeout_none_leaves_client_default() -> None:
-    """Test that FoundryAgent with timeout=None does not call with_options or mutate the client."""
-
-    mock_project = MagicMock()
-    openai_client_mock = MagicMock()
-    openai_client_mock.timeout = 5.0
-    mock_project.get_openai_client.return_value = openai_client_mock
-
-    FoundryAgent(
-        project_client=mock_project,
-        agent_name="test-agent",
-        timeout=None,
-    )
-
-    openai_client_mock.with_options.assert_not_called()
-    assert openai_client_mock.timeout == 5.0
-
-
 def test_raw_foundry_agent_init_rejects_invalid_client_type() -> None:
     """Test that invalid client_type raises TypeError."""
 

python/packages/foundry_hosting/agent_framework_foundry_hosting/_responses.py

@@ -11,7 +11,7 @@ import tempfile
 import threading
 from collections.abc import AsyncIterable, AsyncIterator, Generator, Mapping, Sequence
 from contextlib import AbstractAsyncContextManager, AsyncExitStack, suppress
-from dataclasses import asdict, dataclass, is_dataclass
+from dataclasses import asdict, is_dataclass
 from pathlib import Path
 from typing import Protocol, cast
 
@@ -264,73 +264,28 @@ def _checkpoint_storage_for_context(root: str, context_id: str) -> FileCheckpoin
 
 # Foundry Toolbox Auth integration
 # Consent-URL error code returned by the Foundry MCP gateway when calling `/list`
-CONSENT_ERROR_CODE = -32006
+CONSENT_ERROR_CODE = -32007
 
 
-@dataclass
-class ConsentError:
-    name: str
-    consent_url: str
+def consent_url_from_error(exc: BaseException) -> str | None:
+    """Return the consent URL when ``exc`` wraps a Foundry MCP gateway consent error.
 
-
-def consent_url_from_error(exc: BaseException) -> list[ConsentError] | None:
-    """Return the consent URLs when ``exc`` wraps Foundry MCP gateway consent errors.
+    The Agent Framework MCP layer surfaces gateway consent failures by wrapping the underlying
+    ``McpError`` inside an :class:`AgentFrameworkException` (typically a ``ToolExecutionException``
+    raised from ``MCPStreamableHTTPTool.__aenter__``). This helper inspects ``exc.args`` for a
+    wrapped ``McpError`` whose ``error.code`` is :data:`CONSENT_ERROR_CODE`; when found, the
+    consent link the gateway returned in ``error.message`` is returned. Returns ``None`` for
+    anything else, so callers can do ``if (url := consent_url_from_error(ex)) is None: raise``.
 
     Args:
         exc: The exception to inspect.
 
     Returns:
-        The consent URL(s) extracted from the error, or ``None`` if no consent error was found.
+        The consent URL if ``exc`` wraps a consent ``McpError``, otherwise ``None``.
     """
     inner_exception = next((arg for arg in exc.args if isinstance(arg, McpError)), None)
     if inner_exception is not None and inner_exception.error.code == CONSENT_ERROR_CODE:
-        # Parse the error message
-        # The error message is structured with the following format:
-        # "tools/list failed for 1 tool source(s), succeeded for 0 tool source(s) {"errors":[{"name": ..."
-        # where the second part is a JSON string that can be deserialized into an object with the following shape:
-        # ruff: disable[ERA001]
-        # {
-        #   "errors" : [
-        #       {
-        #           "name": "Name of the MCP tool that requires consent",
-        #           "type" : "mcp",
-        #           "error": {
-        #               "code": "CONSENT_REQUIRED",
-        #               "message": consent_url,
-        #           }
-        #       }
-        #   ]
-        # }
-        # ruff: enable[ERA001]
-        try:
-            consent_errors: list[ConsentError] = []
-            error_message_start = inner_exception.error.message.find("{")
-            if error_message_start == -1:
-                logger.warning("Consent error message does not contain JSON: %s", inner_exception.error.message)
-                return None
-            consent_details_json = inner_exception.error.message[error_message_start:]
-            consent_details = json.loads(consent_details_json)
-            if "errors" not in consent_details or not isinstance(consent_details["errors"], list):
-                logger.warning("Consent error message JSON does not contain 'errors' list: %s", consent_details_json)
-                return None
-            for error in consent_details["errors"]:
-                if (
-                    isinstance(error, dict)
-                    and error.get("type") == "mcp"  # type: ignore
-                    and "error" in error
-                    and isinstance(error["error"], dict)
-                    and error["error"].get("code") == "CONSENT_REQUIRED"  # type: ignore
-                    and "message" in error["error"]
-                ):
-                    consent_url = error["error"]["message"]  # type: ignore
-                    if isinstance(consent_url, str):
-                        consent_errors.append(ConsentError(name=error.get("name", "Unknown"), consent_url=consent_url))  # type: ignore
-                    else:
-                        logger.warning("Consent URL in error message is not a valid URL: %s", consent_url)  # type: ignore
-            if consent_errors:
-                return consent_errors
-        except json.JSONDecodeError:
-            logger.warning("Failed to parse consent details JSON: %s", inner_exception.error.message)
+        return inner_exception.error.message
     return None
 
 
@@ -493,19 +448,18 @@ class ResponsesHostServer(ResponsesAgentServerHost):
         try:
             await self._ensure_agent_ready()
         except AgentFrameworkException as ex:
-            consent_errors = consent_url_from_error(ex)
-            if consent_errors is None:
+            consent_url = consent_url_from_error(ex)
+            if consent_url is None:
                 raise
-            for consent_error in consent_errors:
-                logger.warning("Consent URL for tool '%s': %s", consent_error.name, consent_error.consent_url)
-                oauth_item = OAuthConsentRequestOutputItem(
-                    id=IdGenerator.new_id("oacr"),
-                    consent_link=consent_error.consent_url,
-                    server_label=consent_error.name,
-                )
-                builder = response_event_stream.add_output_item(oauth_item.id)
-                yield builder.emit_added(oauth_item)
-                yield builder.emit_done(oauth_item)
+            logger.warning("OAuth consent required for Foundry MCP gateway.")
+            oauth_item = OAuthConsentRequestOutputItem(
+                id=IdGenerator.new_id("oacr"),
+                consent_link=consent_url,
+                server_label="Foundry Toolbox",
+            )
+            builder = response_event_stream.add_output_item(oauth_item.id)
+            yield builder.emit_added(oauth_item)
+            yield builder.emit_done(oauth_item)
             yield response_event_stream.emit_completed()
             return
 

python/packages/foundry_hosting/pyproject.toml

@@ -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 = "1.0.0a260604"
+version = "1.0.0a260528"
 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>=1.8.0,<2",
+    "agent-framework-core>=1.7.0,<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",

python/packages/foundry_hosting/tests/test_responses.py

@@ -39,7 +39,6 @@ from agent_framework_foundry_hosting import ResponsesHostServer
 from agent_framework_foundry_hosting._responses import (
     _AZURE_RESPONSES_MESSAGE_ROLE_TYPE,  # pyright: ignore[reportPrivateUsage]
     CONSENT_ERROR_CODE,
-    ConsentError,
     FileBasedFunctionApprovalStorage,  # pyright: ignore[reportPrivateUsage]
     InMemoryFunctionApprovalStorage,  # pyright: ignore[reportPrivateUsage]
     _item_to_message,  # pyright: ignore[reportPrivateUsage]
@@ -2119,11 +2118,15 @@ class TestMultiTurnMixedContent:
         assert resp2.json()["status"] == "completed"
 
         second_call_messages = agent.run.call_args_list[1].kwargs["messages"]
-        mcp_call_contents = [c for m in second_call_messages for c in m.contents if c.type == "mcp_server_tool_call"]
+        mcp_call_contents = [
+            c for m in second_call_messages for c in m.contents if c.type == "mcp_server_tool_call"
+        ]
         mcp_result_contents = [
             c for m in second_call_messages for c in m.contents if c.type == "mcp_server_tool_result"
         ]
-        function_result_contents = [c for m in second_call_messages for c in m.contents if c.type == "function_result"]
+        function_result_contents = [
+            c for m in second_call_messages for c in m.contents if c.type == "function_result"
+        ]
 
         assert len(mcp_call_contents) >= 1
         assert len(mcp_result_contents) >= 1
@@ -3261,10 +3264,7 @@ class TestCheckpointContextPathValidation:
 # region Agent lifecycle (lazy entry & OAuth consent surfacing)
 
 
-def _make_consent_error(
-    url: str = "https://consent.example.com/auth",
-    name: str = "Foundry Toolbox",
-) -> Exception:
+def _make_consent_error(url: str = "https://consent.example.com/auth") -> Exception:
     """Build an exception wrapping a Foundry MCP gateway consent error.
 
     Mirrors the real-world wrapping produced by ``MCPStreamableHTTPTool.__aenter__``,
@@ -3272,34 +3272,17 @@ def _make_consent_error(
     ``ToolExecutionException`` (an ``AgentFrameworkException`` subclass) with the
     original error attached via ``inner_exception``. ``consent_url_from_error``
     then finds the wrapped ``McpError`` in ``exc.args``.
-
-    The McpError message uses the structured Foundry MCP gateway format:
-    a human-readable prefix followed by a JSON document describing each
-    failed tool source and its consent URL.
     """
     from agent_framework.exceptions import ToolExecutionException
 
-    payload = json.dumps({
-        "errors": [
-            {
-                "name": name,
-                "type": "mcp",
-                "error": {
-                    "code": "CONSENT_REQUIRED",
-                    "message": url,
-                },
-            }
-        ]
-    })
-    message = f"tools/list failed for 1 tool source(s), succeeded for 0 tool source(s) {payload}"
-    inner = McpError(ErrorData(code=CONSENT_ERROR_CODE, message=message))
+    inner = McpError(ErrorData(code=CONSENT_ERROR_CODE, message=url))
     return ToolExecutionException("MCP consent required", inner_exception=inner)
 
 
 class TestConsentUrlFromError:
     def test_returns_consent_url_when_inner_arg_is_consent_mcp_error(self) -> None:
-        exc = _make_consent_error("https://example.com/consent", name="my-tool")
-        assert consent_url_from_error(exc) == [ConsentError(name="my-tool", consent_url="https://example.com/consent")]
+        exc = _make_consent_error("https://example.com/consent")
+        assert consent_url_from_error(exc) == "https://example.com/consent"
 
     def test_returns_none_when_no_mcp_error_in_args(self) -> None:
         assert consent_url_from_error(Exception("boom")) is None
@@ -3316,13 +3299,6 @@ class TestConsentUrlFromError:
         bare = McpError(ErrorData(code=CONSENT_ERROR_CODE, message="https://x"))
         assert consent_url_from_error(bare) is None
 
-    def test_returns_none_when_message_has_no_json(self) -> None:
-        from agent_framework.exceptions import ToolExecutionException
-
-        inner = McpError(ErrorData(code=CONSENT_ERROR_CODE, message="no json here"))
-        exc = ToolExecutionException("MCP consent required", inner_exception=inner)
-        assert consent_url_from_error(exc) is None
-
 
 class TestAgentLifecycle:
     async def test_agent_entered_lazily_on_first_request(self) -> None:

python/packages/github_copilot/agent_framework_github_copilot/_agent.py

@@ -37,10 +37,9 @@ from agent_framework.exceptions import AgentException
 from agent_framework.observability import AgentTelemetryLayer
 
 try:
-    from copilot import CopilotClient, CopilotSession, RuntimeConnection
-    from copilot.generated.rpc import PermissionDecisionUserNotAvailable
+    from copilot import CopilotClient, CopilotSession, SubprocessConfig
+    from copilot.generated.session_events import PermissionRequest, SessionEvent, SessionEventType
     from copilot.session import MCPServerConfig, PermissionRequestResult, ProviderConfig, SystemMessageConfig
-    from copilot.session_events import PermissionRequest, SessionEvent, SessionEventType
     from copilot.tools import Tool as CopilotTool
     from copilot.tools import ToolInvocation, ToolResult
 except ImportError as _copilot_import_error:
@@ -58,10 +57,8 @@ else:
 DEFAULT_TIMEOUT_SECONDS: float = 60.0
 """Default timeout in seconds for Copilot requests."""
 
-PermissionHandlerType = Callable[
-    [PermissionRequest, dict[str, str]], "PermissionRequestResult | Awaitable[PermissionRequestResult]"
-]
-"""Type for permission request handlers. Supports both sync and async callbacks."""
+PermissionHandlerType = Callable[[PermissionRequest, dict[str, str]], PermissionRequestResult]
+"""Type for permission request handlers."""
 
 
 FunctionApprovalCallback = Callable[[Content], "bool | Awaitable[bool]"]
@@ -124,7 +121,7 @@ def _deny_all_permissions(
     _invocation: dict[str, str],
 ) -> PermissionRequestResult:
     """Default permission handler that denies all requests."""
-    return PermissionDecisionUserNotAvailable()
+    return PermissionRequestResult()
 
 
 class GitHubCopilotSettings(TypedDict, total=False):
@@ -143,9 +140,9 @@ class GitHubCopilotSettings(TypedDict, total=False):
             Can be set via environment variable GITHUB_COPILOT_TIMEOUT.
         log_level: CLI log level.
             Can be set via environment variable GITHUB_COPILOT_LOG_LEVEL.
-        base_directory: Directory where the CLI stores session state, configuration,
+        copilot_home: Directory where the CLI stores session state, configuration,
             and other persistent data. Can be set via environment variable
-            GITHUB_COPILOT_BASE_DIRECTORY. Defaults to ~/.copilot when not set.
+            GITHUB_COPILOT_COPILOT_HOME. Defaults to ~/.copilot when not set.
             Only applicable when the SDK spawns the CLI process (ignored when
             connecting to an external server via a pre-configured client).
     """
@@ -154,7 +151,7 @@ class GitHubCopilotSettings(TypedDict, total=False):
     model: str | None
     timeout: float | None
     log_level: str | None
-    base_directory: str | None
+    copilot_home: str | None
 
 
 class GitHubCopilotOptions(TypedDict, total=False):
@@ -317,7 +314,7 @@ class RawGitHubCopilotAgent(BaseAgent, Generic[OptionsT]):
         provider: ProviderConfig | None = opts.pop("provider", None)
         instruction_directories: list[str] | None = opts.pop("instruction_directories", None)
         on_function_approval: FunctionApprovalCallback | None = opts.pop("on_function_approval", None)
-        base_directory = opts.pop("base_directory", None)
+        copilot_home = opts.pop("copilot_home", None)
 
         self._settings = load_settings(
             GitHubCopilotSettings,
@@ -326,7 +323,7 @@ class RawGitHubCopilotAgent(BaseAgent, Generic[OptionsT]):
             model=model,
             timeout=timeout,
             log_level=log_level,
-            base_directory=base_directory,
+            copilot_home=copilot_home,
             env_file_path=env_file_path,
             env_file_encoding=env_file_encoding,
         )
@@ -365,16 +362,14 @@ class RawGitHubCopilotAgent(BaseAgent, Generic[OptionsT]):
         if self._client is None:
             cli_path = self._settings.get("cli_path") or None
             log_level = self._settings.get("log_level") or None
-            base_directory = self._settings.get("base_directory") or None
+            copilot_home = self._settings.get("copilot_home") or None
 
-            client_kwargs: dict[str, Any] = {}
-            if cli_path:
-                client_kwargs["connection"] = RuntimeConnection.for_stdio(path=cli_path)
+            subprocess_kwargs: dict[str, Any] = {"cli_path": cli_path}
             if log_level:
-                client_kwargs["log_level"] = log_level
-            if base_directory:
-                client_kwargs["base_directory"] = base_directory
-            self._client = CopilotClient(**client_kwargs)
+                subprocess_kwargs["log_level"] = log_level
+            if copilot_home:
+                subprocess_kwargs["copilot_home"] = copilot_home
+            self._client = CopilotClient(SubprocessConfig(**subprocess_kwargs))
 
         try:
             await self._client.start()

python/packages/github_copilot/pyproject.toml

@@ -4,7 +4,7 @@ description = "GitHub Copilot integration for Microsoft Agent Framework."
 authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
 readme = "README.md"
 requires-python = ">=3.10"
-version = "1.0.0rc1"
+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>=1.8.0,<2",
-    "github-copilot-sdk>=1.0.0,<2; python_version >= '3.11'",
+    "agent-framework-core>=1.6.0,<2",
+    "github-copilot-sdk>=1.0.0b2,<=1.0.0b2; python_version >= '3.11'",
 ]
 
 [tool.uv]

python/packages/github_copilot/tests/test_github_copilot_agent.py

@@ -2,7 +2,6 @@
 
 # ruff: noqa: E402
 
-import os
 import unittest.mock
 from datetime import datetime, timezone
 from typing import Any
@@ -21,11 +20,9 @@ from agent_framework import (
     ContextProvider,
     HistoryProvider,
     Message,
-    tool,
 )
 from agent_framework.exceptions import AgentException
-from copilot.session import PermissionHandler
-from copilot.session_events import (
+from copilot.generated.session_events import (
     Data,
     SessionEvent,
     SessionEventType,
@@ -311,27 +308,27 @@ class TestGitHubCopilotAgentLifecycle:
             )
             await agent.start()
 
-            kwargs = MockClient.call_args.kwargs
-            assert kwargs["connection"].path == "/custom/path"
-            assert kwargs["log_level"] == "debug"
+            call_args = MockClient.call_args[0][0]
+            assert call_args.cli_path == "/custom/path"
+            assert call_args.log_level == "debug"
 
-    async def test_start_passes_base_directory_to_client(self) -> None:
-        """Test that base_directory is passed through to CopilotClient."""
+    async def test_start_passes_copilot_home_to_subprocess_config(self) -> None:
+        """Test that copilot_home is passed through to SubprocessConfig."""
         with patch("agent_framework_github_copilot._agent.CopilotClient") as MockClient:
             mock_client = MagicMock()
             mock_client.start = AsyncMock()
             MockClient.return_value = mock_client
 
             agent: GitHubCopilotAgent[GitHubCopilotOptions] = GitHubCopilotAgent(
-                default_options={"base_directory": "/custom/copilot/home"}
+                default_options={"copilot_home": "/custom/copilot/home"}
             )
             await agent.start()
 
-            kwargs = MockClient.call_args.kwargs
-            assert kwargs["base_directory"] == "/custom/copilot/home"
+            call_args = MockClient.call_args[0][0]
+            assert call_args.copilot_home == "/custom/copilot/home"
 
-    async def test_start_base_directory_not_set_when_unspecified(self) -> None:
-        """Test that base_directory is not included in client kwargs when not specified."""
+    async def test_start_copilot_home_not_set_when_unspecified(self) -> None:
+        """Test that copilot_home is not included in SubprocessConfig when not specified."""
         with patch("agent_framework_github_copilot._agent.CopilotClient") as MockClient:
             mock_client = MagicMock()
             mock_client.start = AsyncMock()
@@ -340,14 +337,14 @@ class TestGitHubCopilotAgentLifecycle:
             agent = GitHubCopilotAgent()
             await agent.start()
 
-            kwargs = MockClient.call_args.kwargs
-            assert "base_directory" not in kwargs
+            call_args = MockClient.call_args[0][0]
+            assert call_args.copilot_home is None
 
-    async def test_start_base_directory_from_env_variable(self) -> None:
-        """Test that base_directory can be set via GITHUB_COPILOT_BASE_DIRECTORY env variable."""
+    async def test_start_copilot_home_from_env_variable(self) -> None:
+        """Test that copilot_home can be set via GITHUB_COPILOT_COPILOT_HOME env variable."""
         with (
             patch("agent_framework_github_copilot._agent.CopilotClient") as MockClient,
-            patch.dict("os.environ", {"GITHUB_COPILOT_BASE_DIRECTORY": "/env/copilot/home"}),
+            patch.dict("os.environ", {"GITHUB_COPILOT_COPILOT_HOME": "/env/copilot/home"}),
         ):
             mock_client = MagicMock()
             mock_client.start = AsyncMock()
@@ -356,8 +353,8 @@ class TestGitHubCopilotAgentLifecycle:
             agent = GitHubCopilotAgent()
             await agent.start()
 
-            kwargs = MockClient.call_args.kwargs
-            assert kwargs["base_directory"] == "/env/copilot/home"
+            call_args = MockClient.call_args[0][0]
+            assert call_args.copilot_home == "/env/copilot/home"
 
 
 class TestGitHubCopilotAgentRun:
@@ -1056,11 +1053,11 @@ class TestGitHubCopilotAgentSessionManagement:
         mock_session: MagicMock,
     ) -> None:
         """Test that resumed session config includes tools and permission handler."""
-        from copilot.session import PermissionDecisionApproveOnce, PermissionRequestResult
-        from copilot.session_events import PermissionRequest
+        from copilot.generated.session_events import PermissionRequest
+        from copilot.session import PermissionRequestResult
 
         def my_handler(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
-            return PermissionDecisionApproveOnce()
+            return PermissionRequestResult(kind="approved")
 
         def my_tool(arg: str) -> str:
             """A test tool."""
@@ -1872,15 +1869,6 @@ class TestGitHubCopilotAgentErrorHandling:
 class TestGitHubCopilotAgentPermissions:
     """Test cases for permission handling."""
 
-    def test_deny_all_permissions_returns_user_not_available(self) -> None:
-        """Test that the default deny handler returns PermissionDecisionUserNotAvailable."""
-        from copilot.generated.rpc import PermissionDecisionUserNotAvailable
-
-        from agent_framework_github_copilot._agent import _deny_all_permissions
-
-        result = _deny_all_permissions(MagicMock(), {})
-        assert isinstance(result, PermissionDecisionUserNotAvailable)
-
     def test_no_permission_handler_when_not_provided(self) -> None:
         """Test that no handler is set when on_permission_request is not provided."""
         agent = GitHubCopilotAgent()
@@ -1888,14 +1876,13 @@ class TestGitHubCopilotAgentPermissions:
 
     def test_permission_handler_set_when_provided(self) -> None:
         """Test that a handler is set when on_permission_request is provided."""
-        from copilot.generated.rpc import PermissionDecisionDeniedInteractivelyByUser
-        from copilot.session import PermissionDecisionApproveOnce, PermissionRequestResult
-        from copilot.session_events import PermissionRequest
+        from copilot.generated.session_events import PermissionRequest
+        from copilot.session import PermissionRequestResult
 
         def approve_shell(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
             if request.kind == "shell":
-                return PermissionDecisionApproveOnce()
-            return PermissionDecisionDeniedInteractivelyByUser()
+                return PermissionRequestResult(kind="approved")
+            return PermissionRequestResult(kind="denied-interactively-by-user")
 
         agent: GitHubCopilotAgent[GitHubCopilotOptions] = GitHubCopilotAgent(
             default_options={"on_permission_request": approve_shell}
@@ -1908,14 +1895,13 @@ class TestGitHubCopilotAgentPermissions:
         mock_session: MagicMock,
     ) -> None:
         """Test that session config includes permission handler when provided."""
-        from copilot.generated.rpc import PermissionDecisionDeniedInteractivelyByUser
-        from copilot.session import PermissionDecisionApproveOnce, PermissionRequestResult
-        from copilot.session_events import PermissionRequest
+        from copilot.generated.session_events import PermissionRequest
+        from copilot.session import PermissionRequestResult
 
         def approve_shell_read(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
             if request.kind in ("shell", "read"):
-                return PermissionDecisionApproveOnce()
-            return PermissionDecisionDeniedInteractivelyByUser()
+                return PermissionRequestResult(kind="approved")
+            return PermissionRequestResult(kind="denied-interactively-by-user")
 
         agent: GitHubCopilotAgent[GitHubCopilotOptions] = GitHubCopilotAgent(
             client=mock_client,
@@ -2719,163 +2705,3 @@ class TestGitHubCopilotAgentContextProviders:
         assert call_kwargs.get("tools") is not None
         tool_names = [t.name for t in call_kwargs["tools"]]
         assert "load_skill" in tool_names
-
-
-# ---------------------------------------------------------------------------
-# Integration tests — require COPILOT_GITHUB_TOKEN env var
-# ---------------------------------------------------------------------------
-
-skip_if_copilot_integration_tests_disabled = pytest.mark.skipif(
-    os.getenv("COPILOT_GITHUB_TOKEN", "") == "",
-    reason="No COPILOT_GITHUB_TOKEN provided; skipping integration tests.",
-)
-
-
-@tool(approval_mode="never_require")
-def get_weather(location: str) -> str:
-    """Get the weather for a given location."""
-    return f"The weather in {location} is sunny with a high of 25C."
-
-
-@pytest.mark.flaky
-@pytest.mark.integration
-@skip_if_copilot_integration_tests_disabled
-async def test_integration_run_with_simple_prompt_returns_response() -> None:
-    """Integration test: basic non-streaming response."""
-    agent = GitHubCopilotAgent(
-        instructions="You are a helpful assistant. Keep your answers short.",
-        default_options={"on_permission_request": PermissionHandler.approve_all},
-    )
-
-    async with agent:
-        session = agent.create_session()
-        response = await agent.run("What is 2 + 2? Answer with just the number.", session=session)
-
-        assert response is not None
-        assert len(response.messages) > 0
-        assert "4" in response.text
-
-        if session.service_session_id and agent._client:
-            await agent._client.delete_session(session.service_session_id)
-
-
-@pytest.mark.flaky
-@pytest.mark.integration
-@skip_if_copilot_integration_tests_disabled
-async def test_integration_run_streaming_returns_updates() -> None:
-    """Integration test: streaming response yields updates."""
-    agent = GitHubCopilotAgent(
-        instructions="You are a helpful assistant. Keep your answers short.",
-        default_options={"on_permission_request": PermissionHandler.approve_all},
-    )
-
-    async with agent:
-        session = agent.create_session()
-        updates = []
-        async for chunk in agent.run("Count from 1 to 5.", stream=True, session=session):
-            updates.append(chunk)
-
-        assert len(updates) > 0
-        full_text = "".join(u.text for u in updates if u.text)
-        assert len(full_text) > 0
-
-        if session.service_session_id and agent._client:
-            await agent._client.delete_session(session.service_session_id)
-
-
-@pytest.mark.flaky
-@pytest.mark.integration
-@skip_if_copilot_integration_tests_disabled
-async def test_integration_run_with_function_tool_invokes_tool() -> None:
-    """Integration test: function tool is invoked by the agent."""
-    agent = GitHubCopilotAgent(
-        instructions="You are a helpful weather agent. Use the get_weather tool to answer weather questions.",
-        tools=[get_weather],
-        default_options={"on_permission_request": PermissionHandler.approve_all},
-    )
-
-    async with agent:
-        session = agent.create_session()
-        response = await agent.run("What's the weather like in Seattle?", session=session)
-
-        assert response is not None
-        assert len(response.messages) > 0
-        assert any(word in response.text.lower() for word in ["sunny", "25", "weather", "seattle"])
-
-        if session.service_session_id and agent._client:
-            await agent._client.delete_session(session.service_session_id)
-
-
-@pytest.mark.flaky
-@pytest.mark.integration
-@skip_if_copilot_integration_tests_disabled
-async def test_integration_run_with_session_maintains_context() -> None:
-    """Integration test: session maintains conversation context across turns."""
-    agent = GitHubCopilotAgent(
-        instructions="You are a helpful assistant. Keep your answers short.",
-        default_options={"on_permission_request": PermissionHandler.approve_all},
-    )
-
-    async with agent:
-        session = agent.create_session()
-
-        response1 = await agent.run("My name is Alice.", session=session)
-        assert response1 is not None
-
-        response2 = await agent.run("What is my name?", session=session)
-
-        assert response2 is not None
-        assert "alice" in response2.text.lower()
-
-        if session.service_session_id and agent._client:
-            await agent._client.delete_session(session.service_session_id)
-
-
-@pytest.mark.flaky
-@pytest.mark.integration
-@skip_if_copilot_integration_tests_disabled
-async def test_integration_run_with_session_resume_continues_conversation() -> None:
-    """Integration test: session can be resumed by ID."""
-    agent = GitHubCopilotAgent(
-        instructions="You are a helpful assistant. Keep your answers short.",
-        default_options={"on_permission_request": PermissionHandler.approve_all},
-    )
-
-    async with agent:
-        session1 = agent.create_session()
-        await agent.run("Remember this number: 42.", session=session1)
-
-        session_id = session1.service_session_id
-        assert session_id is not None
-
-        session2 = AgentSession()
-        session2.service_session_id = session_id
-
-        response = await agent.run("What number did I ask you to remember?", session=session2)
-
-        assert response is not None
-        assert "42" in response.text
-
-        if agent._client:
-            await agent._client.delete_session(session_id)
-
-
-@pytest.mark.flaky
-@pytest.mark.integration
-@skip_if_copilot_integration_tests_disabled
-async def test_integration_run_with_shell_permissions_executes_command() -> None:
-    """Integration test: shell commands can be executed with permission handler."""
-    agent = GitHubCopilotAgent(
-        instructions="You are a helpful assistant that can execute shell commands.",
-        default_options={"on_permission_request": PermissionHandler.approve_all},
-    )
-
-    async with agent:
-        session = agent.create_session()
-        response = await agent.run("Run a shell command to print 'hello world'", session=session)
-
-        assert response is not None
-        assert "hello" in response.text.lower()
-
-        if session.service_session_id and agent._client:
-            await agent._client.delete_session(session.service_session_id)

python/packages/mistral/pyproject.toml

@@ -4,7 +4,7 @@ description = "Mistral AI integration for Microsoft Agent Framework."
 authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
 readme = "README.md"
 requires-python = ">=3.10"
-version = "1.0.0a260604"
+version = "1.0.0a260505"
 license-files = ["LICENSE"]
 urls.homepage = "https://learn.microsoft.com/en-us/agent-framework/"
 urls.source = "https://github.com/microsoft/agent-framework/tree/main/python"
@@ -23,7 +23,7 @@ classifiers = [
   "Typing :: Typed",
 ]
 dependencies = [
-    "agent-framework-core>=1.8.0,<2",
+    "agent-framework-core>=1.1.0,<2",
     "mistralai>=2.0.0,<3",
 ]
 

python/packages/openai/agent_framework_openai/_chat_client.py

@@ -385,7 +385,6 @@ class RawOpenAIChatClient(  # type: ignore[misc]
         additional_properties: dict[str, Any] | None = None,
         env_file_path: str | None = None,
         env_file_encoding: str | None = None,
-        timeout: float | None = None,
     ) -> None:
         """Initialize a raw OpenAI Chat client.
 
@@ -407,7 +406,6 @@ class RawOpenAIChatClient(  # type: ignore[misc]
             env_file_path: Optional ``.env`` file that is checked before the process environment
                 for ``OPENAI_*`` values.
             env_file_encoding: Encoding for the ``.env`` file.
-            timeout: Optional timeout in seconds for requests.
         """
         ...
 
@@ -429,7 +427,6 @@ class RawOpenAIChatClient(  # type: ignore[misc]
         additional_properties: dict[str, Any] | None = None,
         env_file_path: str | None = None,
         env_file_encoding: str | None = None,
-        timeout: float | None = None,
     ) -> None:
         """Initialize a raw OpenAI Chat client.
 
@@ -458,7 +455,6 @@ class RawOpenAIChatClient(  # type: ignore[misc]
             env_file_path: Optional ``.env`` file that is checked before process environment
                 variables for ``AZURE_OPENAI_*`` values.
             env_file_encoding: Encoding for the ``.env`` file.
-            timeout: Optional timeout in seconds for requests.
         """
         ...
 
@@ -480,7 +476,6 @@ class RawOpenAIChatClient(  # type: ignore[misc]
         additional_properties: dict[str, Any] | None = None,
         env_file_path: str | None = None,
         env_file_encoding: str | None = None,
-        timeout: float | None = None,
     ) -> None:
         """Initialize a raw OpenAI Chat client.
 
@@ -516,8 +511,6 @@ class RawOpenAIChatClient(  # type: ignore[misc]
                 variables. The same file is used for both ``OPENAI_*`` and ``AZURE_OPENAI_*``
                 lookups.
             env_file_encoding: Encoding for the ``.env`` file.
-            timeout: HTTP timeout in seconds for requests. When not provided, the
-                OpenAI SDK default is used (connect: 5s, total: 600s).
 
         Notes:
             Environment resolution and routing precedence are:
@@ -548,7 +541,6 @@ class RawOpenAIChatClient(  # type: ignore[misc]
             openai_model_fields=("chat_model", "model"),
             azure_model_fields=("chat_model", "model"),
             responses_mode=True,
-            timeout=timeout,
         )
 
         self.client = client
@@ -1462,21 +1454,10 @@ class RawOpenAIChatClient(  # type: ignore[misc]
         Returns:
             The prepared chat messages for a request.
         """
-        drops_reasoning_without_storage = not request_uses_service_side_storage and any(
-            content.type == "text_reasoning" for message in chat_messages for content in message.contents
-        )
-        drop_mcp_call_ids: set[str] = set()
-        if drops_reasoning_without_storage:
-            for message in chat_messages:
-                for content in message.contents:
-                    if content.type == "mcp_server_tool_call" and content.call_id:
-                        drop_mcp_call_ids.add(content.call_id)
-
         list_of_list = [
             self._prepare_message_for_openai(
                 message,
                 request_uses_service_side_storage=request_uses_service_side_storage,
-                drop_mcp_call_ids=drop_mcp_call_ids,
             )
             for message in chat_messages
         ]
@@ -1491,7 +1472,6 @@ class RawOpenAIChatClient(  # type: ignore[misc]
         message: Message,
         *,
         request_uses_service_side_storage: bool = True,
-        drop_mcp_call_ids: set[str] | None = None,
     ) -> list[dict[str, Any]]:
         """Prepare a chat message for the OpenAI Responses API format."""
         all_messages: list[dict[str, Any]] = []
@@ -1511,10 +1491,7 @@ class RawOpenAIChatClient(  # type: ignore[misc]
         # (replays_local_storage) still need stripping when the request also carries a continuation
         # marker, since the server-stored items would otherwise duplicate the inline ones. Without
         # storage, standalone reasoning items are invalid per the API ("reasoning was provided
-        # without its required following item"), so the reasoning branch always drops. When that
-        # happens, `_prepare_messages_for_openai` also drops the paired hosted-MCP IDs across
-        # message boundaries rather than replaying bare MCP items.
-        drop_mcp_call_ids = drop_mcp_call_ids or set()
+        # without its required following item"), so the reasoning branch always drops.
         for content in message.contents:
             match content.type:
                 case "text_reasoning":
@@ -1569,10 +1546,7 @@ class RawOpenAIChatClient(  # type: ignore[misc]
                     # server-side `id`, so under continuation it would duplicate
                     # the prior response's items (#3295). Drop the call here; the
                     # orphan result is dropped by the coalesce step that follows.
-                    #
-                    # Without storage, a reasoning + hosted-MCP pair cannot be replayed
-                    # partially: reasoning is stripped above, and a bare mcp_call is rejected.
-                    if request_uses_service_side_storage or content.call_id in drop_mcp_call_ids:
+                    if request_uses_service_side_storage:
                         continue
                     prepared_mcp = self._prepare_content_for_openai(
                         message.role,

python/packages/openai/agent_framework_openai/_shared.py

@@ -162,7 +162,6 @@ def load_openai_service_settings(
     openai_model_fields: Sequence[OpenAIModelSettingName] = ("model",),
     azure_model_fields: Sequence[OpenAIModelSettingName] = ("model",),
     responses_mode: bool = False,
-    timeout: float | None = None,
 ) -> tuple[dict[str, Any], AsyncOpenAI, bool]:
     """Load OpenAI settings, including Azure OpenAI model aliases.
 
@@ -219,8 +218,6 @@ def load_openai_service_settings(
             }
             if base_url := openai_settings.get("base_url"):
                 client_args["base_url"] = base_url
-            if timeout is not None:
-                client_args["timeout"] = timeout
             return openai_settings, AsyncOpenAI(**client_args), False  # type: ignore[return-value]
         checked_openai = True
     azure_settings = load_settings(
@@ -302,12 +299,8 @@ def load_openai_service_settings(
             openai_args["api_key"] = _ensure_async_token_provider(client_args["azure_ad_token_provider"])
         elif "api_key" in client_args:
             openai_args["api_key"] = client_args["api_key"]
-        if timeout is not None:
-            openai_args["timeout"] = timeout
         return azure_settings, AsyncOpenAI(**openai_args), True  # type: ignore[return-value]
 
-    if timeout is not None:
-        client_args["timeout"] = timeout
     return azure_settings, AsyncAzureOpenAI(**client_args), True  # type: ignore[return-value]
 
 

python/packages/openai/pyproject.toml

@@ -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 = "1.8.0"
+version = "1.7.0"
 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>=1.8.0,<2",
+    "agent-framework-core>=1.7.0,<2",
     "openai>=1.99.0,<3",
 ]
 

python/packages/openai/tests/openai/test_openai_chat_client.py

@@ -36,7 +36,7 @@ from agent_framework.exceptions import (
     ChatClientInvalidRequestException,
     SettingNotFoundError,
 )
-from openai import AsyncOpenAI, BadRequestError
+from openai import BadRequestError
 from openai.types.responses.response_reasoning_item import Summary
 from openai.types.responses.response_reasoning_summary_text_delta_event import (
     ResponseReasoningSummaryTextDeltaEvent,
@@ -55,7 +55,7 @@ from pydantic import BaseModel
 from pytest import param
 
 from agent_framework_openai import OpenAIChatClient
-from agent_framework_openai._chat_client import OPENAI_LOCAL_SHELL_CALL_ITEM_ID_KEY, RawOpenAIChatClient
+from agent_framework_openai._chat_client import OPENAI_LOCAL_SHELL_CALL_ITEM_ID_KEY
 from agent_framework_openai._exceptions import OpenAIContentFilterException
 
 skip_if_openai_integration_tests_disabled = pytest.mark.skipif(
@@ -194,26 +194,6 @@ def test_init_uses_explicit_parameters() -> None:
     assert all(parameter.kind != inspect.Parameter.VAR_KEYWORD for parameter in signature.parameters.values())
 
 
-def test_raw_openai_chat_client_init_uses_explicit_parameters() -> None:
-    signature = inspect.signature(RawOpenAIChatClient.__init__)
-
-    assert "additional_properties" in signature.parameters
-    assert "compaction_strategy" in signature.parameters
-    assert "tokenizer" in signature.parameters
-    assert "timeout" in signature.parameters
-    assert all(parameter.kind != inspect.Parameter.VAR_KEYWORD for parameter in signature.parameters.values())
-
-
-def test_raw_openai_chat_client_accepts_preconfigured_client_with_timeout() -> None:
-    """Test that timeout is accepted without error when async_client is pre-provided."""
-
-    mock_client = MagicMock(spec=AsyncOpenAI)
-    mock_client.timeout = 5.0
-
-    client = RawOpenAIChatClient(async_client=mock_client, timeout=30.0)
-    assert client is not None
-
-
 def test_openai_chat_client_supports_all_tool_protocols() -> None:
     assert isinstance(OpenAIChatClient, SupportsCodeInterpreterTool)
     assert isinstance(OpenAIChatClient, SupportsWebSearchTool)
@@ -5668,79 +5648,6 @@ def test_prepare_messages_for_openai_coalesces_mcp_call_and_result_into_single_i
     assert fco_items == [], f"unexpected orphan function_call_output items: {fco_items}"
 
 
-def test_prepare_messages_for_openai_drops_mcp_call_when_paired_reasoning_is_stripped() -> None:
-    client = OpenAIChatClient(model="test-model", api_key="test-key")
-
-    messages = [
-        Message(
-            role="assistant",
-            contents=[
-                Content.from_text_reasoning(id="rs_abc123", text="Need the MCP server."),
-                Content.from_mcp_server_tool_call(
-                    call_id="mcp_abc123",
-                    tool_name="search",
-                    server_name="api_specs",
-                    arguments='{"q": "cats"}',
-                ),
-            ],
-        ),
-        Message(
-            role="tool",
-            contents=[
-                Content.from_mcp_server_tool_result(
-                    call_id="mcp_abc123",
-                    output=[Content.from_text(text="found 10 cats")],
-                )
-            ],
-        ),
-    ]
-
-    result = client._prepare_messages_for_openai(messages, request_uses_service_side_storage=False)
-
-    types = [item.get("type") for item in result if isinstance(item, dict)]
-    assert "reasoning" not in types
-    assert "mcp_call" not in types
-    assert "function_call_output" not in types
-
-
-def test_prepare_messages_for_openai_drops_mcp_call_across_reasoning_messages() -> None:
-    client = OpenAIChatClient(model="test-model", api_key="test-key")
-
-    messages = [
-        Message(
-            role="assistant",
-            contents=[Content.from_text_reasoning(id="rs_abc123", text="Need a tool call.")],
-        ),
-        Message(
-            role="assistant",
-            contents=[
-                Content.from_mcp_server_tool_call(
-                    call_id="mcp_abc123",
-                    tool_name="search",
-                    server_name="api_specs",
-                    arguments='{"q": "cats"}',
-                )
-            ],
-        ),
-        Message(
-            role="tool",
-            contents=[
-                Content.from_mcp_server_tool_result(
-                    call_id="mcp_abc123",
-                    output=[Content.from_text(text="found 10 cats")],
-                )
-            ],
-        ),
-    ]
-
-    result = client._prepare_messages_for_openai(messages, request_uses_service_side_storage=False)
-
-    types = [item.get("type") for item in result if isinstance(item, dict)]
-    assert "reasoning" not in types
-    assert "mcp_call" not in types
-    assert "function_call_output" not in types
-
-
 def test_prepare_messages_for_openai_drops_orphan_mcp_server_tool_result() -> None:
     """When an mcp_server_tool_result has no matching mcp_server_tool_call in
     the message list, it must be dropped, NOT serialized as a

python/packages/orchestrations/agent_framework_orchestrations/_concurrent.py

@@ -19,7 +19,7 @@ from typing_extensions import Never
 
 from ._orchestration_request_info import AgentApprovalExecutor
 from ._participant_output_config import (
-    UNSET,
+    _MISSING,  # pyright: ignore[reportPrivateUsage]
     _coalesce_output_from,  # pyright: ignore[reportPrivateUsage]
     _coerce_intermediate_output_from,  # pyright: ignore[reportPrivateUsage]
     _ParticipantIntermediateOutputSelection,  # pyright: ignore[reportPrivateUsage]
@@ -213,7 +213,7 @@ class ConcurrentBuilder:
         *,
         participants: Sequence[SupportsAgentRun | Executor],
         checkpoint_storage: CheckpointStorage | None = None,
-        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, UNSET),
+        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, _MISSING),
         intermediate_output_from: _ParticipantIntermediateOutputSelection = None,
     ) -> None:
         """Initialize the ConcurrentBuilder.

python/packages/orchestrations/agent_framework_orchestrations/_group_chat.py

@@ -52,7 +52,7 @@ from ._base_group_chat_orchestrator import (
 from ._orchestration_request_info import AgentApprovalExecutor
 from ._orchestrator_helpers import clean_conversation_for_handoff
 from ._participant_output_config import (
-    UNSET,
+    _MISSING,  # pyright: ignore[reportPrivateUsage]
     _coalesce_output_from,  # pyright: ignore[reportPrivateUsage]
     _coerce_intermediate_output_from,  # pyright: ignore[reportPrivateUsage]
     _ParticipantIntermediateOutputSelection,  # pyright: ignore[reportPrivateUsage]
@@ -626,7 +626,7 @@ class GroupChatBuilder:
         termination_condition: TerminationCondition | None = None,
         max_rounds: int | None = None,
         checkpoint_storage: CheckpointStorage | None = None,
-        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, UNSET),
+        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, _MISSING),
         intermediate_output_from: _ParticipantIntermediateOutputSelection = None,
     ) -> None:
         """Initialize the GroupChatBuilder.

python/packages/orchestrations/agent_framework_orchestrations/_handoff.py

@@ -54,7 +54,7 @@ from agent_framework._workflows._workflow_context import WorkflowContext
 from ._base_group_chat_orchestrator import TerminationCondition
 from ._orchestrator_helpers import clean_conversation_for_handoff
 from ._participant_output_config import (
-    UNSET,
+    _MISSING,  # pyright: ignore[reportPrivateUsage]
     _coalesce_output_from,  # pyright: ignore[reportPrivateUsage]
     _coerce_intermediate_output_from,  # pyright: ignore[reportPrivateUsage]
     _ParticipantIntermediateOutputSelection,  # pyright: ignore[reportPrivateUsage]
@@ -597,7 +597,7 @@ class HandoffBuilder:
         description: str | None = None,
         checkpoint_storage: CheckpointStorage | None = None,
         termination_condition: TerminationCondition | None = None,
-        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, UNSET),
+        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, _MISSING),
         intermediate_output_from: _ParticipantIntermediateOutputSelection = None,
     ) -> None:
         r"""Initialize a HandoffBuilder for creating conversational handoff workflows.

python/packages/orchestrations/agent_framework_orchestrations/_magentic.py

@@ -28,7 +28,7 @@ from agent_framework._workflows._request_info_mixin import response_handler
 from agent_framework._workflows._workflow import Workflow
 from agent_framework._workflows._workflow_builder import WorkflowBuilder
 from agent_framework._workflows._workflow_context import WorkflowContext
-from typing_extensions import Never, Sentinel
+from typing_extensions import Never
 
 from ._base_group_chat_orchestrator import (
     BaseGroupChatOrchestrator,
@@ -39,7 +39,7 @@ from ._base_group_chat_orchestrator import (
     ParticipantRegistry,
 )
 from ._participant_output_config import (
-    UNSET,
+    _MISSING,  # pyright: ignore[reportPrivateUsage]
     _coalesce_output_from,  # pyright: ignore[reportPrivateUsage]
     _coerce_intermediate_output_from,  # pyright: ignore[reportPrivateUsage]
     _ParticipantIntermediateOutputSelection,  # pyright: ignore[reportPrivateUsage]
@@ -1411,13 +1411,13 @@ class MagenticBuilder:
         task_ledger_plan_update_prompt: str | None = None,
         progress_ledger_prompt: str | None = None,
         final_answer_prompt: str | None = None,
-        max_stall_count: int | Sentinel = UNSET,
+        max_stall_count: int = 3,
         max_reset_count: int | None = None,
         max_round_count: int | None = None,
         # Existing params
         enable_plan_review: bool = False,
         checkpoint_storage: CheckpointStorage | None = None,
-        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, UNSET),
+        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, _MISSING),
         intermediate_output_from: _ParticipantIntermediateOutputSelection = None,
     ) -> None:
         """Initialize the Magentic workflow builder.
@@ -1621,7 +1621,7 @@ class MagenticBuilder:
         progress_ledger_prompt: str | None = None,
         final_answer_prompt: str | None = None,
         # Limits
-        max_stall_count: int | Sentinel = UNSET,
+        max_stall_count: int = 3,
         max_reset_count: int | None = None,
         max_round_count: int | None = None,
     ) -> None:
@@ -1656,10 +1656,8 @@ class MagenticBuilder:
                 "Exactly one of manager, manager_agent, manager_factory, or manager_agent_factory must be provided."
             )
 
-        resolved_max_stall_count: int = 3 if max_stall_count is UNSET else cast(int, max_stall_count)
-
         def _log_warning_if_constructor_args_provided() -> None:
-            if max_stall_count is not UNSET or any(
+            if any(
                 arg is not None
                 for arg in [
                     task_ledger,
@@ -1670,6 +1668,7 @@ class MagenticBuilder:
                     task_ledger_plan_update_prompt,
                     progress_ledger_prompt,
                     final_answer_prompt,
+                    max_stall_count,
                     max_reset_count,
                     max_round_count,
                 ]
@@ -1690,7 +1689,7 @@ class MagenticBuilder:
                 task_ledger_plan_update_prompt=task_ledger_plan_update_prompt,
                 progress_ledger_prompt=progress_ledger_prompt,
                 final_answer_prompt=final_answer_prompt,
-                max_stall_count=resolved_max_stall_count,
+                max_stall_count=max_stall_count,
                 max_reset_count=max_reset_count,
                 max_round_count=max_round_count,
             )
@@ -1708,7 +1707,7 @@ class MagenticBuilder:
                 "task_ledger_plan_update_prompt": task_ledger_plan_update_prompt,
                 "progress_ledger_prompt": progress_ledger_prompt,
                 "final_answer_prompt": final_answer_prompt,
-                "max_stall_count": resolved_max_stall_count,
+                "max_stall_count": max_stall_count,
                 "max_reset_count": max_reset_count,
                 "max_round_count": max_round_count,
             }

python/packages/orchestrations/agent_framework_orchestrations/_participant_output_config.py

@@ -8,9 +8,8 @@ from typing import Any, Literal
 from agent_framework import SupportsAgentRun
 from agent_framework._workflows._agent_utils import resolve_agent_id
 from agent_framework._workflows._executor import Executor
-from typing_extensions import Sentinel
 
-UNSET = Sentinel("UNSET")
+_MISSING = object()
 _ALL_OUTPUTS: Literal["all"] = "all"
 _ALL_OTHER_OUTPUTS: Literal["all_other"] = "all_other"
 _ParticipantOutputSpecifier = str | SupportsAgentRun | Executor
@@ -21,10 +20,10 @@ _WorkflowExecutorSpecifier = Executor | SupportsAgentRun
 
 def _coalesce_output_from(  # pyright: ignore[reportUnusedFunction]
     *,
-    output_from: Any = UNSET,
+    output_from: Any = _MISSING,
 ) -> _ParticipantOutputSelection:
     """Resolve orchestration output selection to ``output_from``."""
-    if output_from is not UNSET:
+    if output_from is not _MISSING:
         return _coerce_output_from(output_from)
     return None
 

python/packages/orchestrations/agent_framework_orchestrations/_sequential.py

@@ -33,7 +33,7 @@ from agent_framework._workflows._workflow_context import WorkflowContext
 
 from ._orchestration_request_info import AgentApprovalExecutor
 from ._participant_output_config import (
-    UNSET,
+    _MISSING,  # pyright: ignore[reportPrivateUsage]
     _coalesce_output_from,  # pyright: ignore[reportPrivateUsage]
     _coerce_intermediate_output_from,  # pyright: ignore[reportPrivateUsage]
     _ParticipantIntermediateOutputSelection,  # pyright: ignore[reportPrivateUsage]
@@ -99,7 +99,7 @@ class SequentialBuilder:
         participants: Sequence[SupportsAgentRun | Executor],
         checkpoint_storage: CheckpointStorage | None = None,
         chain_only_agent_responses: bool = False,
-        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, UNSET),
+        output_from: Sequence[_ParticipantOutputSpecifier] | Literal["all"] | None = cast(Any, _MISSING),
         intermediate_output_from: _ParticipantIntermediateOutputSelection = None,
     ) -> None:
         """Initialize the SequentialBuilder.

python/packages/orchestrations/pyproject.toml

@@ -4,7 +4,7 @@ description = "Orchestration patterns for Microsoft Agent Framework. Includes Se
 authors = [{ name = "Microsoft", email = "af-support@microsoft.com"}]
 readme = "README.md"
 requires-python = ">=3.10"
-version = "1.0.0rc3"
+version = "1.0.0rc2"
 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>=1.8.0,<2",
+    "agent-framework-core>=1.6.0,<2",
 ]
 
 [tool.uv]

python/packages/orchestrations/tests/test_magentic.py

@@ -1,6 +1,5 @@
 # Copyright (c) Microsoft. All rights reserved.
 
-import logging
 import sys
 from collections.abc import AsyncIterable, Awaitable, Sequence
 from dataclasses import dataclass
@@ -988,33 +987,6 @@ def test_magentic_builder_requires_exactly_one_manager_option():
         MagenticBuilder(participants=[agent], manager=manager, manager_factory=manager_factory)
 
 
-def test_magentic_with_custom_manager_does_not_warn_without_standard_manager_options(caplog: Any) -> None:
-    caplog.set_level(logging.WARNING, logger="agent_framework_orchestrations._magentic")
-
-    MagenticBuilder(participants=[StubAgent("agentA", "reply")], manager=FakeManager())
-
-    assert "Custom manager provided; all other manager arguments will be ignored." not in caplog.text
-
-
-def test_magentic_with_custom_manager_factory_does_not_warn_without_standard_manager_options(caplog: Any) -> None:
-    caplog.set_level(logging.WARNING, logger="agent_framework_orchestrations._magentic")
-
-    def manager_factory() -> MagenticManagerBase:
-        return FakeManager()
-
-    MagenticBuilder(participants=[StubAgent("agentA", "reply")], manager_factory=manager_factory)
-
-    assert "Custom manager provided; all other manager arguments will be ignored." not in caplog.text
-
-
-def test_magentic_with_custom_manager_warns_when_standard_manager_option_is_provided(caplog: Any) -> None:
-    caplog.set_level(logging.WARNING, logger="agent_framework_orchestrations._magentic")
-
-    MagenticBuilder(participants=[StubAgent("agentA", "reply")], manager=FakeManager(), max_stall_count=3)
-
-    assert "Custom manager provided; all other manager arguments will be ignored." in caplog.text
-
-
 async def test_magentic_with_manager_factory():
     """Test workflow creation using manager_factory."""
     factory_call_count = 0
@@ -1065,20 +1037,6 @@ async def test_magentic_with_agent_factory():
     assert event_count > 0
 
 
-def test_magentic_agent_factory_uses_default_max_stall_count() -> None:
-    def agent_factory() -> SupportsAgentRun:
-        return cast(SupportsAgentRun, StubManagerAgent())
-
-    participant = StubAgent("agentA", "reply from agentA")
-    workflow = MagenticBuilder(participants=[participant], manager_agent_factory=agent_factory).build()
-
-    orchestrator = next(e for e in workflow.executors.values() if isinstance(e, MagenticOrchestrator))
-    manager = orchestrator._manager  # type: ignore[reportPrivateUsage]
-
-    assert isinstance(manager, StandardMagenticManager)
-    assert manager.max_stall_count == 3
-
-
 async def test_magentic_manager_factory_reusable_builder():
     """Test that the builder can be reused to build multiple workflows with manager factory."""
     factory_call_count = 0

python/pyproject.toml

@@ -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 = "1.8.0"
+version = "1.7.0"
 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]==1.8.0",
+    "agent-framework-core[all]==1.7.0",
 ]
 
 [dependency-groups]

python/samples/02-agents/compaction/README.md

@@ -5,8 +5,7 @@ This folder demonstrates context compaction patterns introduced by ADR-0019.
 ## Files
 
 - `basics.py` — builds a local message list and applies each built-in strategy one at a time.
-- `summarization.py` — runs `SummarizationStrategy` directly with a real summarizing chat client.
-- `advanced.py` — composes multiple strategies with `TokenBudgetComposedStrategy`, including a real summarizer and tool-call groups.
+- `advanced.py` — composes multiple strategies with `TokenBudgetComposedStrategy`.
 - `agent_client_overrides.py` — shows client defaults, agent-level overrides, and per-run compaction overrides.
 - `custom.py` — defines a custom strategy implementing the `CompactionStrategy` protocol.
 - `tiktoken_tokenizer.py` — shows a `TokenizerProtocol` implementation backed by `tiktoken`.
@@ -16,8 +15,7 @@ Run samples with:
 
 ```bash
 uv run samples/02-agents/compaction/basics.py
-uv run samples/02-agents/compaction/summarization.py  # requires OPENAI_API_KEY
-uv run samples/02-agents/compaction/advanced.py  # requires OPENAI_API_KEY
+uv run samples/02-agents/compaction/advanced.py
 uv run samples/02-agents/compaction/agent_client_overrides.py
 uv run samples/02-agents/compaction/custom.py
 uv run samples/02-agents/compaction/tiktoken_tokenizer.py

python/samples/02-agents/compaction/advanced.py

@@ -1,14 +1,11 @@
 # Copyright (c) Microsoft. All rights reserved.
 
 import asyncio
-from typing import Any, cast
+from typing import Any
 
 from agent_framework import (
-    GROUP_ANNOTATION_KEY,
-    GROUP_TOKEN_COUNT_KEY,
-    SUMMARY_OF_MESSAGE_IDS_KEY,
     CharacterEstimatorTokenizer,
-    Content,
+    ChatResponse,
     Message,
     SelectiveToolCallCompactionStrategy,
     SlidingWindowStrategy,
@@ -18,48 +15,36 @@ from agent_framework import (
     apply_compaction,
     included_token_count,
 )
-from agent_framework.openai import OpenAIChatClient
-from dotenv import load_dotenv
 
-load_dotenv()
-
-"""This sample demonstrates composed in-run compaction under a token budget.
-
-A long, tool-using conversation is compacted with a single
-``TokenBudgetComposedStrategy`` that runs three strategies in order until the
-included-token count fits the budget:
-
-1. ``SelectiveToolCallCompactionStrategy`` — drop older tool-call groups
-   (assistant ``function_call`` + ``tool`` result messages) that are expensive
-   and rarely needed verbatim once acted upon.
-2. ``SummarizationStrategy`` — use a *real* chat client to summarize the oldest
-   remaining turns into a single linked summary message.
-3. ``SlidingWindowStrategy`` — as a final guard, keep only the most recent
-   groups if the budget is still exceeded.
+"""This sample demonstrates composed in-run compaction with a token budget.
 
 Key components:
-- TokenBudgetComposedStrategy with ordered, escalating strategies
-- A real OpenAIChatClient used as the summarizer (not a stub)
-- Tool-call groups in the history so tool-call compaction is meaningful
-- Token accounting before/after via a TokenizerProtocol
-
-Run with:
-    uv run samples/02-agents/compaction/advanced.py  # requires OPENAI_API_KEY
+- TokenBudgetComposedStrategy
+- Sequential strategy composition
+- Summarization with a SupportsChatGetResponse-compatible summarizer client
 """
 
 
-def _build_long_history() -> list[Message]:
-    """Build a long, tool-using migration conversation to create token pressure."""
-    history: list[Message] = [
-        Message(role="system", contents=["You are a migration copilot that plans and executes database migrations."]),
-    ]
+class BudgetSummaryClient:
+    async def get_response(
+        self,
+        messages: list[Message],
+        *,
+        stream: bool = False,
+        options: dict[str, Any] | None = None,
+        **kwargs: Any,
+    ) -> ChatResponse:
+        summary_text = f"Budget summary generated from {len(messages)} prompt messages."
+        return ChatResponse(messages=[Message(role="assistant", contents=[summary_text])])
 
-    # A few verbose planning turns to build up token pressure.
-    for i in range(1, 5):
+
+def _build_long_history() -> list[Message]:
+    history = [Message(role="system", contents=["You are a migration copilot."])]
+    for i in range(1, 8):
         history.append(
             Message(
                 role="user",
-                contents=[f"Iteration {i}: capture migration requirements, constraints, and edge cases in detail."],
+                contents=[f"Iteration {i}: capture migration requirements and edge cases."],
             )
         )
         history.append(
@@ -67,62 +52,17 @@ def _build_long_history() -> list[Message]:
                 role="assistant",
                 contents=[
                     (
-                        f"Iteration {i}: produced a detailed plan covering dependencies, rollback guidance, data "
-                        "backfill, and a full testing matrix. This response is intentionally verbose to add pressure."
+                        f"Iteration {i}: detailed plan with dependencies, rollback guidance, and testing details. "
+                        "This sentence is intentionally long to create token pressure."
                     )
                 ],
             )
         )
-
-    # A tool-call group: the assistant inspects the schema via a tool.
-    history.append(
-        Message(
-            role="assistant",
-            contents=[Content.from_function_call(call_id="call_1", name="inspect_schema", arguments='{"db":"legacy"}')],
-        )
-    )
-    history.append(
-        Message(
-            role="tool",
-            contents=[Content.from_function_result(call_id="call_1", result="tables: users, orders, invoices, events")],
-        )
-    )
-    history.append(Message(role="assistant", contents=["Schema inspection found four core tables to migrate."]))
-
-    # The most recent turn — this should survive compaction verbatim.
-    history.append(Message(role="user", contents=["What is the safest order to migrate these tables?"]))
-    history.append(
-        Message(
-            role="assistant",
-            contents=["Migrate reference tables (users) first, then orders, then invoices, and events last."],
-        )
-    )
     return history
 
 
-def _annotation(message: Message) -> dict[str, Any] | None:
-    annotation = message.additional_properties.get(GROUP_ANNOTATION_KEY)
-    return cast("dict[str, Any]", annotation) if isinstance(annotation, dict) else None
-
-
-def _token_count(message: Message) -> int | None:
-    annotation = _annotation(message)
-    return annotation.get(GROUP_TOKEN_COUNT_KEY) if annotation else None
-
-
-def _relation(message: Message) -> str:
-    """Describe how a projected message relates to the original messages."""
-    annotation = _annotation(message)
-    if annotation is None:
-        return ""
-    summarizes = annotation.get(SUMMARY_OF_MESSAGE_IDS_KEY)
-    if summarizes:
-        return f" <- summary of {summarizes}"
-    return ""
-
-
 async def main() -> None:
-    # 1. Build synthetic history representing long-running, tool-using growth.
+    # 1. Build synthetic history representing long-running in-run growth.
     messages = _build_long_history()
 
     # 2. Configure tokenizer and measure token count before compaction.
@@ -130,35 +70,22 @@ async def main() -> None:
     annotate_message_groups(messages, tokenizer=tokenizer)
     budget_before = included_token_count(messages)
 
-    print("Before compaction message set:")
-    for msg in messages:
-        text_preview = msg.text[:80] if msg.text else "<non-text>"
-        print(f"- [{msg.role}] {text_preview} ({msg.message_id}, {_token_count(msg)} tokens)")
-    print()
-
-    # 3. Create a real summarizer client. SummarizationStrategy only requires a
-    #    SupportsChatGetResponse-compatible client.
-    summarizer = OpenAIChatClient(model="gpt-4o-mini")
-
-    # 4. Configure the composed strategy stack. Strategies run in order and the
-    #    composed strategy stops as soon as the included-token budget is met.
-    #    The budget is set high enough that the generated summary fits within it:
-    #    a tighter budget would trip the composed fallback, which excludes the
-    #    oldest group first (the summary) once the included set exceeds the
-    #    budget. SlidingWindowStrategy remains as a recency safety net for longer
-    #    histories; for this sample summarization alone reaches budget, so the
-    #    window does not need to fire.
+    # 3. Configure composed strategy stack.
     composed = TokenBudgetComposedStrategy(
-        token_budget=400,
+        token_budget=200,
         tokenizer=tokenizer,
         strategies=[
             SelectiveToolCallCompactionStrategy(keep_last_tool_call_groups=0),
-            SummarizationStrategy(client=summarizer, target_count=3, threshold=2),
+            SummarizationStrategy(
+                client=BudgetSummaryClient(),
+                target_count=3,
+                threshold=3,
+            ),
             SlidingWindowStrategy(keep_last_groups=4),
         ],
     )
 
-    # 5. Apply compaction and inspect the budget result.
+    # 4. Apply compaction and inspect the budget result.
     projected = await apply_compaction(messages, strategy=composed, tokenizer=tokenizer)
     budget_after = included_token_count(messages)
 
@@ -168,44 +95,23 @@ async def main() -> None:
     print("Projected roles:", [m.role for m in projected])
     print("Projected messages with token counts:")
     for msg in projected:
+        group = msg.additional_properties.get("_group")
+        token_count = group.get("token_count") if isinstance(group, dict) else None
         text_preview = msg.text[:80] if msg.text else "<non-text>"
-        print(f"- [{msg.role}] {text_preview} ({msg.message_id}, {_token_count(msg)} tokens){_relation(msg)}")
-
-    # 6. Surface the model-generated summary, if summarization fired.
-    for msg in messages:
-        annotation = _annotation(msg)
-        if annotation and annotation.get(SUMMARY_OF_MESSAGE_IDS_KEY):
-            print("\nGenerated summary:")
-            print(f"  {msg.text}")
-            print(f"  summarizes: {annotation.get(SUMMARY_OF_MESSAGE_IDS_KEY)}")
+        print(f"- [{msg.role}] {text_preview} ({token_count} tokens)")
 
 
 if __name__ == "__main__":
     asyncio.run(main())
 
 """
-Sample output (summary text and token counts vary because the summary is generated by the model):
-
-Before compaction message set:
-- [system] You are a migration copilot that plans and executes database migrations. (msg_0, 46 tokens)
-- [user] Iteration 1: capture migration requirements, constraints, and edge cases in deta (msg_1, 48 tokens)
-- [assistant] Iteration 1: produced a detailed plan covering dependencies, rollback guidance,  (msg_2, 73 tokens)
-...
-- [user] What is the safest order to migrate these tables? (msg_12, 40 tokens)
-- [assistant] Migrate reference tables (users) first, then orders, then invoices, and events l (msg_13, 50 tokens)
-
-Projected messages after compaction: 5
-Included token count before compaction: 757
-Included token count after compaction: 274
-Projected roles: ['system', 'assistant', 'assistant', 'user', 'assistant']
+Sample output:
+Projected messages after compaction: 3
+Included token count before compaction: 793
+Included token count after compaction: 144
+Projected roles: ['system', 'user', 'assistant']
 Projected messages with token counts:
-- [system] You are a migration copilot that plans and executes database migrations. (msg_0, 46 tokens)
-- [assistant] Across four planning turns the user and assistant... (summary_14, 96 tokens) <- summary of [msg_1..8]
-- [assistant] Schema inspection found four core tables to migrate. (msg_11, 42 tokens)
-- [user] What is the safest order to migrate these tables? (msg_12, 40 tokens)
-- [assistant] Migrate reference tables (users) first, then orders, then invoices, and events l (msg_13, 50 tokens)
-
-Generated summary:
-  Across four planning turns the user and assistant defined the migration requirements...
-  summarizes: ['msg_1', 'msg_2', 'msg_3', 'msg_4', 'msg_5', 'msg_6', 'msg_7', 'msg_8']
+- [system] You are a migration copilot. (35 tokens)
+- [user] Iteration 7: capture migration requirements and edge cases. (43 tokens)
+- [assistant] Iteration 7: detailed plan with dependencies, rollback guidance, and testing det (66 tokens)
 """

python/samples/02-agents/compaction/summarization.py

@@ -1,159 +0,0 @@
-# Copyright (c) Microsoft. All rights reserved.
-
-import asyncio
-from typing import Any, cast
-
-from agent_framework import (
-    GROUP_ANNOTATION_KEY,
-    SUMMARIZED_BY_SUMMARY_ID_KEY,
-    SUMMARY_OF_MESSAGE_IDS_KEY,
-    Message,
-    SummarizationStrategy,
-    apply_compaction,
-)
-from agent_framework.openai import OpenAIChatClient
-from dotenv import load_dotenv
-
-load_dotenv()
-
-"""This sample demonstrates the SummarizationStrategy directly.
-
-Unlike SlidingWindow/Truncation strategies that simply drop older groups,
-``SummarizationStrategy`` calls a real chat client to *summarize* the oldest
-message groups, replaces them with a single linked summary message, and keeps
-the most recent turns verbatim. This preserves long-range context (decisions,
-goals, unresolved items) while bounding the prompt size.
-
-Key components:
-- SummarizationStrategy with a real OpenAIChatClient summarizer
-- ``apply_compaction`` to run the strategy over a message list
-- Bidirectional summary trace metadata (summary -> originals, original -> summary)
-
-Run with:
-    uv run samples/02-agents/compaction/summarization.py  # requires OPENAI_API_KEY
-"""
-
-
-def _annotation(message: Message) -> dict[str, Any] | None:
-    annotation = message.additional_properties.get(GROUP_ANNOTATION_KEY)
-    return cast("dict[str, Any]", annotation) if isinstance(annotation, dict) else None
-
-
-def _build_history() -> list[Message]:
-    """Build a multi-turn conversation long enough to trigger summarization."""
-    return [
-        Message(role="system", contents=["You are a project planning assistant."]),
-        Message(role="user", contents=["We are migrating a monolith to microservices. Where do we start?"]),
-        Message(
-            role="assistant",
-            contents=["Start by mapping bounded contexts and identifying the highest-churn modules to extract first."],
-        ),
-        Message(role="user", contents=["The billing module changes most often. What are the risks of extracting it?"]),
-        Message(
-            role="assistant",
-            contents=["Main risks: distributed transactions, invoices-table ownership, and latency on hot paths."],
-        ),
-        Message(role="user", contents=["How should we handle the shared invoices table?"]),
-        Message(
-            role="assistant",
-            contents=["Use the strangler-fig pattern: dual-write during transition, then make billing the owner."],
-        ),
-        Message(role="user", contents=["What is the most recent decision we made?"]),
-        Message(role="assistant", contents=["We decided to extract billing first using the strangler-fig pattern."]),
-    ]
-
-
-def _print_messages(label: str, messages: list[Message]) -> None:
-    print(f"\n--- {label} ---")
-    print(f"Message count: {len(messages)}")
-    for index, message in enumerate(messages, start=1):
-        text = message.text or ", ".join(content.type for content in message.contents)
-        print(f"{index:02d}. [{message.role}] {text[:90]}")
-
-
-async def main() -> None:
-    # 1. Create a real summarizing client. SummarizationStrategy only requires a
-    #    SupportsChatGetResponse-compatible client, so any chat client works.
-    summarizer = OpenAIChatClient(model="gpt-4o-mini")
-
-    # 2. Build a conversation and show it before compaction.
-    messages = _build_history()
-    _print_messages("Before compaction", messages)
-
-    # 3. Configure the strategy. It triggers once the included non-system message
-    #    count exceeds ``target_count + threshold`` (here 4 + 2 = 6), summarizing
-    #    the oldest groups down toward ``target_count`` while keeping recent turns.
-    strategy = SummarizationStrategy(
-        client=summarizer,
-        target_count=4,
-        threshold=2,
-    )
-
-    # 4. Apply the strategy. The oldest groups are summarized into a single
-    #    assistant message; the projected list is what the model would receive.
-    projected = await apply_compaction(messages, strategy=strategy)
-    _print_messages("After compaction (SummarizationStrategy)", projected)
-
-    # 5. Inspect the generated summary and its bidirectional trace metadata.
-    print("\n--- Summary trace ---")
-    for message in messages:
-        annotation = _annotation(message)
-        if annotation is None:
-            continue
-        summarizes = annotation.get(SUMMARY_OF_MESSAGE_IDS_KEY)
-        if summarizes:
-            print(f"Generated summary ({message.message_id}):")
-            print(f"  {message.text}")
-            print(f"  summarizes original ids: {summarizes}")
-    summarized_by: dict[str | None, Any] = {}
-    for message in messages:
-        annotation = _annotation(message)
-        if annotation is None:
-            continue
-        summary_id = annotation.get(SUMMARIZED_BY_SUMMARY_ID_KEY)
-        if summary_id:
-            summarized_by[message.message_id] = summary_id
-    if summarized_by:
-        print("Originals replaced by the summary:")
-        for original_id, summary_id in summarized_by.items():
-            print(f"  {original_id} -> {summary_id}")
-
-
-if __name__ == "__main__":
-    asyncio.run(main())
-
-"""
-Sample output (summary text varies because it is generated by the model):
-
---- Before compaction ---
-Message count: 9
-01. [system] You are a project planning assistant.
-02. [user] We are migrating a monolith to microservices. Where do we start?
-03. [assistant] Start by mapping bounded contexts and identifying the highest-churn modules to ex
-04. [user] The billing module changes most often. What are the risks of extracting it?
-05. [assistant] Main risks: distributed transactions, data ownership of the invoices table, and lat
-06. [user] How should we handle the shared invoices table?
-07. [assistant] Use the strangler-fig pattern: dual-write during transition, then make billing the
-08. [user] What is the most recent decision we made?
-09. [assistant] We decided to extract billing first using the strangler-fig pattern.
-
---- After compaction (SummarizationStrategy) ---
-Message count: 6
-01. [system] You are a project planning assistant.
-02. [assistant] The user is migrating a monolith to microservices and decided to extract the billin
-03. [user] How should we handle the shared invoices table?
-04. [assistant] Use the strangler-fig pattern: dual-write during transition, then make billing the
-05. [user] What is the most recent decision we made?
-06. [assistant] We decided to extract billing first using the strangler-fig pattern.
-
---- Summary trace ---
-Generated summary (summary_9):
-  The user is migrating a monolith to microservices and decided to extract the billing module first...
-  summarizes original ids: ['msg_1', 'msg_2', 'msg_3', 'msg_4', 'msg_5']
-Originals replaced by the summary:
-  msg_1 -> summary_9
-  msg_2 -> summary_9
-  msg_3 -> summary_9
-  msg_4 -> summary_9
-  msg_5 -> summary_9
-"""

python/samples/02-agents/mcp/README.md

@@ -13,9 +13,12 @@ The Model Context Protocol (MCP) is an open standard for connecting AI agents to
 | **Agent as MCP Server** | [`agent_as_mcp_server.py`](agent_as_mcp_server.py) | Shows how to expose an Agent Framework agent as an MCP server that other AI applications can connect to |
 | **API Key Authentication** | [`mcp_api_key_auth.py`](mcp_api_key_auth.py) | Demonstrates API key authentication with MCP servers using `header_provider`, runtime invocation kwargs, and a command-line API key argument |
 | **GitHub Integration with PAT** | [`mcp_github_pat.py`](mcp_github_pat.py) | Demonstrates connecting to GitHub's MCP server using Personal Access Token (PAT) authentication |
+| **Long-Running Task** | [`mcp_long_running_task.py`](mcp_long_running_task.py) | Demonstrates transparent SEP-2663 long-running task handling for MCP tools that advertise `taskSupport=required`. Self-spawns a stdio MCP child server |
 
 ## Prerequisites
 
+Most samples in this folder use OpenAI:
+
 - `OPENAI_API_KEY` environment variable
 - `OPENAI_CHAT_MODEL` environment variable
 
@@ -23,3 +26,8 @@ Run `mcp_api_key_auth.py` with the MCP API key as the first command-line argumen
 
 For `mcp_github_pat.py`:
 - `GITHUB_PAT` - Your GitHub Personal Access Token (create at https://github.com/settings/tokens)
+
+For `mcp_long_running_task.py` (uses Azure OpenAI via Entra-ID):
+- Run `az login` once
+- `AZURE_OPENAI_ENDPOINT` - your Azure OpenAI resource endpoint, e.g. `https://<resource>.openai.azure.com/`
+- `AZURE_OPENAI_CHAT_MODEL` (or `AZURE_OPENAI_MODEL`) - the deployment name (e.g. `gpt-4o-mini`)

python/samples/02-agents/mcp/mcp_long_running_task.py

@@ -0,0 +1,181 @@
+# Copyright (c) Microsoft. All rights reserved.
+
+"""
+MCP Long-Running Task (SEP-2663) Example
+
+Demonstrates that ``MCPStdioTool`` transparently drives the MCP long-running
+task lifecycle for tools that advertise ``execution.taskSupport == "required"``.
+The agent observes a single function-call result; the framework handles the
+``tools/call`` → ``tasks/get`` (polled) → ``tasks/result`` sequence in the
+background.
+
+Run it as a single file. The script doubles as both the client and the stdio
+MCP child server (the child branch is selected via ``--server``):
+
+    python mcp_long_running_task.py
+
+Requirements:
+- Azure CLI sign-in (``az login``) — used for Entra-ID auth against Azure OpenAI.
+- ``AZURE_OPENAI_ENDPOINT`` — your Azure OpenAI resource endpoint, e.g.
+  ``https://<resource>.openai.azure.com/``.
+- ``AZURE_OPENAI_CHAT_MODEL`` (or ``AZURE_OPENAI_MODEL``) — the deployment name,
+  e.g. ``gpt-4o-mini``.
+
+This sample uses the lower-level ``mcp.server.lowlevel.Server`` so it can:
+1. Advertise a tool with ``execution=ToolExecution(taskSupport="required")``.
+2. Enable the SDK's experimental task support for the ``tasks/*`` lifecycle.
+"""
+
+import asyncio
+import sys
+from datetime import timedelta
+from typing import Any
+
+from agent_framework import Agent, MCPStdioTool, MCPTaskOptions
+from agent_framework.openai import OpenAIChatClient
+from azure.identity import AzureCliCredential
+from dotenv import load_dotenv
+
+load_dotenv()
+
+
+# ---------------------------------------------------------------------------
+# MCP stdio server (child-process branch)
+# ---------------------------------------------------------------------------
+
+
+async def _run_server() -> None:
+    """Run a minimal stdio MCP server exposing one long-running tool."""
+    import mcp.types as types
+    from mcp.server.lowlevel import Server
+    from mcp.server.stdio import stdio_server
+
+    server: Server[Any, Any] = Server("mcp-long-running-task-demo")
+    # Auto-registers handlers for tasks/get, tasks/result, tasks/cancel, tasks/list
+    # backed by an in-memory store.
+    server.experimental.enable_tasks()
+
+    @server.list_tools()
+    async def _list_tools() -> list[types.Tool]:  # pyright: ignore[reportUnusedFunction]
+        return [
+            types.Tool(
+                name="slow_summary",
+                description=(
+                    "Produces a short summary of the supplied text after simulating several seconds of expensive work."
+                ),
+                inputSchema={
+                    "type": "object",
+                    "properties": {
+                        "text": {
+                            "type": "string",
+                            "description": "Text to summarize.",
+                        }
+                    },
+                    "required": ["text"],
+                },
+                # Advertise that this tool MUST be invoked via the task lifecycle.
+                execution=types.ToolExecution(taskSupport="required"),
+            )
+        ]
+
+    @server.call_tool()
+    async def _call_tool(name: str, arguments: dict[str, Any]) -> Any:  # pyright: ignore[reportUnusedFunction]
+        if name != "slow_summary":
+            raise ValueError(f"Unknown tool: {name}")
+
+        ctx = server.request_context
+
+        async def _work(task: Any) -> types.CallToolResult:
+            await task.update_status("Thinking...")
+            await asyncio.sleep(15.0)
+            text: str = (arguments.get("text") or "").strip()
+            words = text.split()
+            preview = " ".join(words[:6]) + ("..." if len(words) > 6 else "")
+            summary = (
+                f"Summarized {len(words)} word(s). First few words: '{preview}'."
+                if words
+                else "No input text was provided."
+            )
+            return types.CallToolResult(
+                content=[types.TextContent(type="text", text=summary)],
+                isError=False,
+            )
+
+        if not ctx.experimental.is_task:
+            # Client invoked the tool without task augmentation. Return a hard
+            # error so a misconfigured client surfaces the problem clearly.
+            return types.CallToolResult(
+                content=[
+                    types.TextContent(
+                        type="text",
+                        text="'slow_summary' must be invoked as a task.",
+                    )
+                ],
+                isError=True,
+            )
+
+        return await ctx.experimental.run_task(_work)
+
+    async with stdio_server() as (read_stream, write_stream):
+        await server.run(read_stream, write_stream, server.create_initialization_options())
+
+
+# ---------------------------------------------------------------------------
+# Agent client (default branch)
+# ---------------------------------------------------------------------------
+
+
+async def _run_client() -> None:
+    mcp_tool = MCPStdioTool(
+        name="LongRunningDemo",
+        description="Demo MCP server exposing a tool that advertises taskSupport=required.",
+        command=sys.executable,
+        args=[__file__, "--server"],
+        # Optional: cap individual tasks at two minutes. The server may apply its
+        # own default if this is omitted.
+        task_options=MCPTaskOptions(default_ttl=timedelta(minutes=2)),
+    )
+
+    async with Agent(
+        client=OpenAIChatClient(credential=AzureCliCredential()),
+        name="LROAgent",
+        instructions=(
+            "You are a helpful assistant. Use the slow_summary tool when the user "
+            "asks for a summary. Wait for the result and present it directly."
+        ),
+        tools=mcp_tool,
+    ) as agent:
+        prompt = (
+            "Please summarize the following text using your slow_summary tool: "
+            "'The Model Context Protocol lets language models talk to external "
+            "tools and resources through a small JSON-RPC surface.'"
+        )
+
+        print("=== run() ===")
+        print(f"User: {prompt}")
+        response = await agent.run(prompt)
+        print(f"Agent: {response.text}\n")
+
+        print("=== run(stream=True) ===")
+        print(f"User: {prompt}")
+        print("Agent: ", end="", flush=True)
+        async for update in agent.run(prompt, stream=True):
+            if update.text:
+                print(update.text, end="", flush=True)
+        print()
+
+
+# ---------------------------------------------------------------------------
+# Entry point
+# ---------------------------------------------------------------------------
+
+
+def main() -> None:
+    if len(sys.argv) > 1 and sys.argv[1] == "--server":
+        asyncio.run(_run_server())
+        return
+    asyncio.run(_run_client())
+
+
+if __name__ == "__main__":
+    main()

python/samples/02-agents/providers/github_copilot/README.md

@@ -23,7 +23,7 @@ The following environment variables can be configured:
 | `GITHUB_COPILOT_MODEL` | Model to use (e.g., "gpt-5", "claude-sonnet-4") | Server default |
 | `GITHUB_COPILOT_TIMEOUT` | Request timeout in seconds | `60` |
 | `GITHUB_COPILOT_LOG_LEVEL` | CLI log level | `info` |
-| `GITHUB_COPILOT_BASE_DIRECTORY` | Directory for CLI session state and config | `~/.copilot` |
+| `GITHUB_COPILOT_COPILOT_HOME` | Directory for CLI session state and config | `~/.copilot` |
 
 ## Observability
 

python/samples/02-agents/providers/github_copilot/github_copilot_basic.py

@@ -19,7 +19,8 @@ from typing import Annotated
 
 from agent_framework import tool
 from agent_framework.github import GitHubCopilotAgent
-from copilot.session import PermissionHandler
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import PermissionRequestResult
 from dotenv import load_dotenv
 from pydantic import Field
 
@@ -27,6 +28,19 @@ from pydantic import Field
 load_dotenv()
 
 
+def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
+    """Permission handler that prompts the user for approval."""
+    print(f"\n[Permission Request: {request.kind}]")
+
+    if request.full_command_text is not None:
+        print(f"  Command: {request.full_command_text}")
+
+    response = input("Approve? (y/n): ").strip().lower()
+    if response in ("y", "yes"):
+        return PermissionRequestResult(kind="approved")
+    return PermissionRequestResult(kind="denied-interactively-by-user")
+
+
 # NOTE: approval_mode="never_require" is for sample brevity. Use "always_require" in production;
 # see samples/02-agents/tools/function_tool_with_approval.py
 # and samples/02-agents/tools/function_tool_with_approval_and_sessions.py.
@@ -46,7 +60,7 @@ async def non_streaming_example() -> None:
     agent = GitHubCopilotAgent(
         instructions="You are a helpful weather agent.",
         tools=[get_weather],
-        default_options={"on_permission_request": PermissionHandler.approve_all},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent:
@@ -63,7 +77,7 @@ async def streaming_example() -> None:
     agent = GitHubCopilotAgent(
         instructions="You are a helpful weather agent.",
         tools=[get_weather],
-        default_options={"on_permission_request": PermissionHandler.approve_all},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent:
@@ -83,7 +97,7 @@ async def runtime_options_example() -> None:
     agent = GitHubCopilotAgent(
         instructions="Always respond in exactly 3 words.",
         tools=[get_weather],
-        default_options={"on_permission_request": PermissionHandler.approve_all},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent:

python/samples/02-agents/providers/github_copilot/github_copilot_with_file_operations.py

@@ -4,7 +4,8 @@
 GitHub Copilot Agent with File Operation Permissions
 
 This sample demonstrates how to enable file read and write operations with GitHubCopilotAgent.
-By providing a permission handler, the agent can read from and write to files on the filesystem.
+By providing a permission handler that approves "read" and/or "write" requests, the agent can
+read from and write to files on the filesystem.
 
 SECURITY NOTE: Only enable file permissions when you trust the agent's actions.
 - "read" allows the agent to read any accessible file
@@ -14,18 +15,21 @@ SECURITY NOTE: Only enable file permissions when you trust the agent's actions.
 import asyncio
 
 from agent_framework.github import GitHubCopilotAgent
-from copilot.generated.rpc import PermissionDecisionDeniedInteractivelyByUser
-from copilot.session import PermissionHandler, PermissionRequestResult
-from copilot.session_events import PermissionRequest
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import PermissionRequestResult
 
 
-async def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
+def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
     """Permission handler that prompts the user for approval."""
     print(f"\n[Permission Request: {request.kind}]")
-    response = (await asyncio.to_thread(input, "Approve? (y/n): ")).strip().lower()
+
+    if request.path is not None:
+        print(f"  Path: {request.path}")
+
+    response = input("Approve? (y/n): ").strip().lower()
     if response in ("y", "yes"):
-        return PermissionHandler.approve_all(request, context)
-    return PermissionDecisionDeniedInteractivelyByUser()
+        return PermissionRequestResult(kind="approved")
+    return PermissionRequestResult(kind="denied-interactively-by-user")
 
 
 async def main() -> None:

python/samples/02-agents/providers/github_copilot/github_copilot_with_function_approval.py

@@ -32,7 +32,6 @@ from typing import Annotated
 
 from agent_framework import Content, tool
 from agent_framework.github import GitHubCopilotAgent
-from copilot.session import PermissionHandler
 from dotenv import load_dotenv
 
 load_dotenv()
@@ -49,42 +48,37 @@ def get_weather_detail(location: Annotated[str, "The city and state, e.g. San Fr
     )
 
 
-async def prompt_for_approval(call: Content) -> bool:
-    """Async approval callback that prompts the user interactively.
+def prompt_for_approval(call: Content) -> bool:
+    """Synchronous approval prompt.
 
     The callback receives a ``FunctionCallContent`` so the operator can review
     the tool name and arguments before deciding. Returning ``True`` allows the
     call; returning ``False`` denies it and a tool-error is returned to the
     model.
-
-    Uses ``asyncio.to_thread`` so the event loop is not blocked by ``input()``.
     """
-    print(f"\n  [Function Approval Request]\n  Tool: {call.name}\n  Arguments: {call.arguments}")
-    response = (await asyncio.to_thread(input, "  Approve this tool call? (y/n): ")).strip().lower()
+    print(f"\n[Function Approval Request]\n  Tool: {call.name}\n  Arguments: {call.arguments}")
+    response = input("Approve this tool call? (y/n): ").strip().lower()
     return response in ("y", "yes")
 
 
-def auto_approve(call: Content) -> bool:
-    """Synchronous approval callback that always approves.
+async def prompt_for_approval_async(call: Content) -> bool:
+    """Async approval prompt.
 
-    Use a sync callback for simple, non-blocking decisions that don't require
-    I/O (e.g. checking an allow-list of tool names).
+    Use an async callback when approval requires I/O (e.g. an HTTP call to a
+    review service or queueing the request to a UI). ``input()`` is wrapped
+    with ``asyncio.to_thread`` so the event loop is not blocked.
     """
-    print(f"\n  [Function Approval Request]\n  Tool: {call.name}\n  Arguments: {call.arguments}")
-    print("  -> Auto-approved")
-    return True
+    print(f"\n[Function Approval Request - async]\n  Tool: {call.name}\n  Arguments: {call.arguments}")
+    response = await asyncio.to_thread(input, "Approve this tool call? (y/n): ")
+    return response.strip().lower() in ("y", "yes")
 
 
-async def run_with_interactive_callback() -> None:
-    """Demonstrates an interactive approval prompt before tool execution."""
-    print("\n=== GitHub Copilot Agent: interactive approval callback ===")
+async def run_with_sync_callback() -> None:
+    print("\n=== GitHub Copilot Agent: synchronous approval callback ===")
     agent = GitHubCopilotAgent(
         instructions="You are a helpful weather assistant.",
         tools=[get_weather_detail],
-        default_options={
-            "on_function_approval": prompt_for_approval,
-            "on_permission_request": PermissionHandler.approve_all,
-        },
+        default_options={"on_function_approval": prompt_for_approval},
     )
     async with agent:
         query = "Give me the detailed weather for Seattle."
@@ -93,16 +87,12 @@ async def run_with_interactive_callback() -> None:
         print(f"Agent: {result}")
 
 
-async def run_with_auto_approve_callback() -> None:
-    """Demonstrates a synchronous callback that always approves."""
-    print("\n=== GitHub Copilot Agent: synchronous auto-approve callback ===")
+async def run_with_async_callback() -> None:
+    print("\n=== GitHub Copilot Agent: asynchronous approval callback ===")
     agent = GitHubCopilotAgent(
         instructions="You are a helpful weather assistant.",
         tools=[get_weather_detail],
-        default_options={
-            "on_function_approval": auto_approve,
-            "on_permission_request": PermissionHandler.approve_all,
-        },
+        default_options={"on_function_approval": prompt_for_approval_async},
     )
     async with agent:
         query = "Give me the detailed weather for Tokyo."
@@ -122,7 +112,6 @@ async def run_without_callback() -> None:
     agent = GitHubCopilotAgent(
         instructions="You are a helpful weather assistant.",
         tools=[get_weather_detail],
-        default_options={"on_permission_request": PermissionHandler.approve_all},
     )
     async with agent:
         query = "Give me the detailed weather for Paris."
@@ -133,8 +122,8 @@ async def run_without_callback() -> None:
 
 async def main() -> None:
     print("=== GitHub Copilot Agent: Function approval enforcement ===")
-    await run_with_interactive_callback()
-    await run_with_auto_approve_callback()
+    await run_with_sync_callback()
+    await run_with_async_callback()
     await run_without_callback()
 
 

python/samples/02-agents/providers/github_copilot/github_copilot_with_instruction_directories.py

@@ -22,13 +22,24 @@ import asyncio
 from pathlib import Path
 
 from agent_framework.github import GitHubCopilotAgent
-from copilot.session import PermissionHandler
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import PermissionRequestResult
 from dotenv import load_dotenv
 
 # Load environment variables from .env file
 load_dotenv()
 
 
+def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
+    """Permission handler that prompts the user for approval."""
+    print(f"\n[Permission Request: {request.kind}]")
+
+    response = input("Approve? (y/n): ").strip().lower()
+    if response in ("y", "yes"):
+        return PermissionRequestResult(kind="approved")
+    return PermissionRequestResult(kind="denied-interactively-by-user")
+
+
 async def default_instructions_example() -> None:
     """Example of pointing the agent at project-specific instruction directories."""
     print("=== Instruction Directories (Default) ===\n")
@@ -47,7 +58,7 @@ async def default_instructions_example() -> None:
     agent = GitHubCopilotAgent(
         instructions="You are a helpful coding assistant.",
         default_options={
-            "on_permission_request": PermissionHandler.approve_all,
+            "on_permission_request": prompt_permission,
             "instruction_directories": instruction_dirs,
         },
     )
@@ -68,7 +79,7 @@ async def runtime_override_example() -> None:
     agent = GitHubCopilotAgent(
         instructions="You are a helpful assistant.",
         default_options={
-            "on_permission_request": PermissionHandler.approve_all,
+            "on_permission_request": prompt_permission,
             "instruction_directories": ["/team/shared/instructions"],
         },
     )

python/samples/02-agents/providers/github_copilot/github_copilot_with_mcp.py

@@ -15,13 +15,24 @@ of MCP-related actions.
 import asyncio
 
 from agent_framework.github import GitHubCopilotAgent
-from copilot.session import MCPServerConfig, PermissionHandler
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import MCPServerConfig, PermissionRequestResult
 from dotenv import load_dotenv
 
 # Load environment variables from .env file
 load_dotenv()
 
 
+def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
+    """Permission handler that prompts the user for approval."""
+    print(f"\n[Permission Request: {request.kind}]")
+
+    response = input("Approve? (y/n): ").strip().lower()
+    if response in ("y", "yes"):
+        return PermissionRequestResult(kind="approved")
+    return PermissionRequestResult(kind="denied-interactively-by-user")
+
+
 async def main() -> None:
     print("=== GitHub Copilot Agent with MCP Servers ===\n")
 
@@ -45,7 +56,7 @@ async def main() -> None:
     agent = GitHubCopilotAgent(
         instructions="You are a helpful assistant with access to the local filesystem and Microsoft Learn.",
         default_options={
-            "on_permission_request": PermissionHandler.approve_all,
+            "on_permission_request": prompt_permission,
             "mcp_servers": mcp_servers,
         },
     )

python/samples/02-agents/providers/github_copilot/github_copilot_with_multiple_permissions.py

@@ -3,8 +3,9 @@
 """
 GitHub Copilot Agent with Multiple Permissions
 
-This sample demonstrates how multiple permission types are requested when GitHubCopilotAgent
-performs complex tasks that require different capabilities.
+This sample demonstrates how to enable multiple permission types with GitHubCopilotAgent.
+By combining different permission kinds in the handler, the agent can perform complex tasks
+that require multiple capabilities.
 
 Available permission kinds:
 - "shell": Execute shell commands
@@ -20,14 +21,23 @@ More permissions mean more potential for unintended actions.
 import asyncio
 
 from agent_framework.github import GitHubCopilotAgent
-from copilot.session import PermissionHandler, PermissionRequestResult
-from copilot.session_events import PermissionRequest
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import PermissionRequestResult
 
 
-def approve_and_log(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
-    """Permission handler that auto-approves and logs each permission kind."""
-    print(f"  [Permission: {request.kind}]", flush=True)
-    return PermissionHandler.approve_all(request, context)
+def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
+    """Permission handler that prompts the user for approval."""
+    print(f"\n[Permission Request: {request.kind}]")
+
+    if request.full_command_text is not None:
+        print(f"  Command: {request.full_command_text}")
+    if request.path is not None:
+        print(f"  Path: {request.path}")
+
+    response = input("Approve? (y/n): ").strip().lower()
+    if response in ("y", "yes"):
+        return PermissionRequestResult(kind="approved")
+    return PermissionRequestResult(kind="denied-interactively-by-user")
 
 
 async def main() -> None:
@@ -35,14 +45,14 @@ async def main() -> None:
 
     agent = GitHubCopilotAgent(
         instructions="You are a helpful development assistant that can read, write files and run commands.",
-        default_options={"on_permission_request": approve_and_log},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent:
         query = "List the first 3 Python files, then read the first one and create a summary in summary.txt"
-        print(f"User: {query}\n")
+        print(f"User: {query}")
         result = await agent.run(query)
-        print(f"\nAgent: {result}\n")
+        print(f"Agent: {result}\n")
 
 
 if __name__ == "__main__":

python/samples/02-agents/providers/github_copilot/github_copilot_with_session.py

@@ -14,10 +14,24 @@ from typing import Annotated
 
 from agent_framework import tool
 from agent_framework.github import GitHubCopilotAgent
-from copilot.session import PermissionHandler
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import PermissionRequestResult
 from pydantic import Field
 
 
+def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
+    """Permission handler that prompts the user for approval."""
+    print(f"\n[Permission Request: {request.kind}]")
+
+    if request.full_command_text is not None:
+        print(f"  Command: {request.full_command_text}")
+
+    response = input("Approve? (y/n): ").strip().lower()
+    if response in ("y", "yes"):
+        return PermissionRequestResult(kind="approved")
+    return PermissionRequestResult(kind="denied-interactively-by-user")
+
+
 # NOTE: approval_mode="never_require" is for sample brevity. Use "always_require" in production;
 # see samples/02-agents/tools/function_tool_with_approval.py
 # and samples/02-agents/tools/function_tool_with_approval_and_sessions.py.
@@ -37,7 +51,7 @@ async def example_with_automatic_session_creation() -> None:
     agent = GitHubCopilotAgent(
         instructions="You are a helpful weather agent.",
         tools=[get_weather],
-        default_options={"on_permission_request": PermissionHandler.approve_all},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent:
@@ -62,7 +76,7 @@ async def example_with_session_persistence() -> None:
     agent = GitHubCopilotAgent(
         instructions="You are a helpful weather agent.",
         tools=[get_weather],
-        default_options={"on_permission_request": PermissionHandler.approve_all},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent:
@@ -99,7 +113,7 @@ async def example_with_existing_session_id() -> None:
     agent1 = GitHubCopilotAgent(
         instructions="You are a helpful weather agent.",
         tools=[get_weather],
-        default_options={"on_permission_request": PermissionHandler.approve_all},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent1:
@@ -121,7 +135,7 @@ async def example_with_existing_session_id() -> None:
         agent2 = GitHubCopilotAgent(
             instructions="You are a helpful weather agent.",
             tools=[get_weather],
-            default_options={"on_permission_request": PermissionHandler.approve_all},
+            default_options={"on_permission_request": prompt_permission},
         )
 
         async with agent2:

python/samples/02-agents/providers/github_copilot/github_copilot_with_shell.py

@@ -14,20 +14,21 @@ Shell commands have full access to your system within the permissions of the run
 import asyncio
 
 from agent_framework.github import GitHubCopilotAgent
-from copilot.generated.rpc import PermissionDecisionUserNotAvailable
-from copilot.session import PermissionHandler, PermissionRequestResult
-from copilot.session_events import PermissionRequest
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import PermissionRequestResult
 
 
-def approve_and_log(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
-    """Permission handler that approves only shell commands and logs them."""
-    if request.kind == "shell":
-        print(f"\n  [Permission: {request.kind}]", flush=True)
-        command = getattr(request, "full_command_text", None)
-        if command is not None:
-            print(f"  Command: {command}", flush=True)
-        return PermissionHandler.approve_all(request, context)
-    return PermissionDecisionUserNotAvailable()
+def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
+    """Permission handler that prompts the user for approval."""
+    print(f"\n[Permission Request: {request.kind}]")
+
+    if request.full_command_text is not None:
+        print(f"  Command: {request.full_command_text}")
+
+    response = input("Approve? (y/n): ").strip().lower()
+    if response in ("y", "yes"):
+        return PermissionRequestResult(kind="approved")
+    return PermissionRequestResult(kind="denied-interactively-by-user")
 
 
 async def main() -> None:
@@ -35,14 +36,14 @@ async def main() -> None:
 
     agent = GitHubCopilotAgent(
         instructions="You are a helpful assistant that can execute shell commands.",
-        default_options={"on_permission_request": approve_and_log},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent:
         query = "List the first 3 Python files in the current directory"
         print(f"User: {query}")
         result = await agent.run(query)
-        print(f"\nAgent: {result}\n")
+        print(f"Agent: {result}\n")
 
 
 if __name__ == "__main__":

python/samples/02-agents/providers/github_copilot/github_copilot_with_url.py

@@ -14,20 +14,21 @@ URL fetching allows the agent to access any URL accessible from your network.
 import asyncio
 
 from agent_framework.github import GitHubCopilotAgent
-from copilot.generated.rpc import PermissionDecisionUserNotAvailable
-from copilot.session import PermissionHandler, PermissionRequestResult
-from copilot.session_events import PermissionRequest
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import PermissionRequestResult
 
 
-def approve_and_log(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
-    """Permission handler that approves only URL requests and logs them."""
-    if request.kind == "url":
-        print(f"\n  [Permission: {request.kind}]", flush=True)
-        url = getattr(request, "url", None)
-        if url is not None:
-            print(f"  URL: {url}", flush=True)
-        return PermissionHandler.approve_all(request, context)
-    return PermissionDecisionUserNotAvailable()
+def prompt_permission(request: PermissionRequest, context: dict[str, str]) -> PermissionRequestResult:
+    """Permission handler that prompts the user for approval."""
+    print(f"\n[Permission Request: {request.kind}]")
+
+    if request.url is not None:
+        print(f"  URL: {request.url}")
+
+    response = input("Approve? (y/n): ").strip().lower()
+    if response in ("y", "yes"):
+        return PermissionRequestResult(kind="approved")
+    return PermissionRequestResult(kind="denied-interactively-by-user")
 
 
 async def main() -> None:
@@ -35,14 +36,14 @@ async def main() -> None:
 
     agent = GitHubCopilotAgent(
         instructions="You are a helpful assistant that can fetch and summarize web content.",
-        default_options={"on_permission_request": approve_and_log},
+        default_options={"on_permission_request": prompt_permission},
     )
 
     async with agent:
         query = "Fetch https://learn.microsoft.com/agent-framework/tutorials/quick-start and summarize its contents"
         print(f"User: {query}")
         result = await agent.run(query)
-        print(f"\nAgent: {result}\n")
+        print(f"Agent: {result}\n")
 
 
 if __name__ == "__main__":

python/samples/04-hosting/foundry-hosted-agents/responses/04_foundry_toolbox/README.md

@@ -20,7 +20,7 @@ You can connect to MCP servers in Foundry Toolbox that use different authenticat
 - **Agent identity authentication**: The tool requires an agent identity token to authenticate. Sample MCP server: `https://{foundry-resource-name}.cognitiveservices.azure.com/language/mcp?api-version=2025-11-15-preview` (Azure Language MCP server) with agent identity for authentication.
 - **Entra Pass-through authentication**: The tool requires an Entra pass-through token to authenticate. Sample MCP server: Microsoft Outlook MCP server with Entra pass-through for authentication.
 
-> Definitions of these authentication methods can be found in the [agent.manifest.yaml](agent.manifest.yaml) file in this sample. The GitHub MCP connection defaults to using a PAT for authentication in this sample, but you can switch to OAuth2 by changing the `project_connection_id` field in the `agent.manifest.yaml` file and following the instructions in the comments.
+> Definitions of these authentication methods can be found in the [agent.manifest.yaml](agent.manifest.yaml) file in this sample.
 
 There are also Non-MCP tools in the toolbox that support different authentication methods. Learn more at the [Foundry sample repository](https://github.com/microsoft-foundry/foundry-samples/blob/main/samples/python/hosted-agents/SUPPORTED_TOOLBOX_SCENARIOS.md).
 

python/samples/04-hosting/foundry-hosted-agents/responses/04_foundry_toolbox/agent.manifest.yaml

@@ -18,92 +18,92 @@ template:
     - name: AZURE_AI_MODEL_DEPLOYMENT_NAME
       value: "{{AZURE_AI_MODEL_DEPLOYMENT_NAME}}"
     - name: TOOLBOX_NAME
-      value: "agent-tools"
-parameters:
-  properties:
-    - name: mcp_endpoint
-      # `azd ai agent init -m` will prompt for this value when initializing the agent manifest
-      secret: false
-      description: URL of the public MCP server (e.g. https://gitmcp.io/Azure/azure-rest-api-specs) that does not require authentication
-    - name: github_pat
-      # `azd ai agent init -m` will prompt for this value when initializing the agent manifest.
-      # Only needed when the GitHub MCP connection is configured to use the `github-mcp-pat-conn`
-      # PAT-based connection below; if you use the `github-mcp-oauth-conn` OAuth2 connection
-      # instead, you can leave this empty.
-      secret: true
-      description: GitHub Personal Access Token used to authenticate with the GitHub MCP server (only needed when using the PAT connection; press Enter if using OAuth2 instead)
-  # - name: language_mcp_entra_audience
-  #   secret: false
-  #   description: Entra ID audience for the Azure Language MCP server (e.g. https://cognitiveservices.azure.com/)
-  # - name: language_mcp_target_url
-  #   secret: false
-  #   description: URL of the Azure Language MCP server that accepts agent identity tokens (e.g. https://{foundry-resource-name}.cognitiveservices.azure.com/language/mcp?api-version=2025-11-15-preview)
-  # - name: outlook_mail_entra_audience
-  #   secret: false
-  #   description: Entra ID audience for the Outlook Mail MCP server
-  # - name: outlook_mail_entra_mcp_target
-  #   secret: false
-  #   description: URL of the Outlook Mail MCP server that accepts user Entra tokens
+      value: "agent-tools-2"
+# parameters:
+#   properties:
+# - name: mcp_endpoint
+#   # `azd ai agent init -m` will prompt for this value when initializing the agent manifest
+#   secret: false
+#   description: URL of the public MCP server (e.g. https://gitmcp.io/Azure/azure-rest-api-specs) that does not require authentication
+# - name: github_pat
+#   # `azd ai agent init -m` will prompt for this value when initializing the agent manifest.
+#   # Only needed when the GitHub MCP connection is configured to use the `github-mcp-pat-conn`
+#   # PAT-based connection below; if you use the `github-mcp-oauth-conn` OAuth2 connection
+#   # instead, you can leave this empty.
+#   secret: true
+#   description: GitHub Personal Access Token used to authenticate with the GitHub MCP server (only needed when using the PAT connection; press Enter if using OAuth2 instead)
+# - name: language_mcp_entra_audience
+#   secret: false
+#   description: Entra ID audience for the Azure Language MCP server (e.g. https://cognitiveservices.azure.com/)
+# - name: language_mcp_target_url
+#   secret: false
+#   description: URL of the Azure Language MCP server that accepts agent identity tokens (e.g. https://{foundry-resource-name}.cognitiveservices.azure.com/language/mcp?api-version=2025-11-15-preview)
+# - name: outlook_mail_entra_audience
+#   secret: false
+#   description: Entra ID audience for the Outlook Mail MCP server
+# - name: outlook_mail_entra_mcp_target
+#   secret: false
+#   description: URL of the Outlook Mail MCP server that accepts user Entra tokens
 resources:
   - kind: model
     id: gpt-4.1-mini
     name: AZURE_AI_MODEL_DEPLOYMENT_NAME
-  - kind: connection
-    # A connection that uses a GitHub Personal Access Token (PAT) to authenticate with the GitHub MCP server
-    name: github-mcp-pat-conn
-    category: RemoteTool
-    authType: CustomKeys
-    target: https://api.githubcopilot.com/mcp
-    credentials:
-      type: CustomKeys
-      keys:
-        Authorization: "Bearer {{ github_pat }}"
-  - kind: connection
-    # A connection that uses OAuth2 to authenticate with the GitHub MCP server
-    name: github-mcp-oauth-conn
-    category: RemoteTool
-    authType: OAuth2
-    target: https://api.githubcopilot.com/mcp
-    connectorName: foundrygithubmcp
-    credentials:
-      type: OAuth2
-      clientId: managed
-      clientSecret: managed
+  # - kind: connection
+  #   # A connection that uses a GitHub Personal Access Token (PAT) to authenticate with the GitHub MCP server
+  #   name: github-mcp-pat-conn
+  #   category: RemoteTool
+  #   authType: CustomKeys
+  #   target: https://api.githubcopilot.com/mcp
+  #   credentials:
+  #     type: CustomKeys
+  #     keys:
+  #       Authorization: "Bearer {{ github_pat }}"
+  # - kind: connection
+  #   # A connection that uses OAuth2 to authenticate with the GitHub MCP server
+  #   name: github-mcp-oauth-conn
+  #   category: RemoteTool
+  #   authType: OAuth2
+  #   target: https://api.githubcopilot.com/mcp
+  #   connectorName: foundrygithubmcp
+  #   credentials:
+  #     type: OAuth2
+  #     clientId: managed
+  #     clientSecret: managed
   # - kind: connection
   #   name: language-mcp-conn
   #   category: RemoteTool
   #   authType: AgenticIdentity
   #   audience: "{{ language_mcp_entra_audience }}"
   #   target: "{{ language_mcp_target_url }}"
-  # - kind: connection
-  #   name: outlook-mail-conn
-  #   category: RemoteTool
-  #   authType: UserEntraToken
-  #   audience: "{{ outlook_mail_entra_audience }}"
-  #   target: "{{ outlook_mail_entra_mcp_target }}"
-  - kind: toolbox
-    name: agent-tools
-    tools:
-      - type: web_search
-        name: web_search
-      - type: code_interpreter
-        name: code_interpreter
-      - type: mcp
-        # This MCP tool doesn't require authentication
-        server_label: noauth_mcp
-        server_url: "{{ mcp_endpoint }}"
-        require_approval: "never"
-      - type: mcp
-        # This MCP tool uses the GitHub MCP server with a PAT for authentication or OAuth2
-        server_label: github
-        project_connection_id: github-mcp-pat-conn # use `github-mcp-oauth-conn` for OAuth2 authentication
-        require_approval: "never"
-      # - type: mcp
-      #   # This MCP tool uses the Azure Language MCP server with agent identity for authentication
-      #   server_label: language-mcp
-      #   project_connection_id: language-mcp-conn
-      #   require_approval: "never"
-      # - type: mcp
-      #   server_label: outlook-mail
-      #   project_connection_id: outlook-mail-conn
-      #   require_approval: "never"
+  # # - kind: connection
+  # #   name: outlook-mail-conn
+  # #   category: RemoteTool
+  # #   authType: UserEntraToken
+  # #   audience: "{{ outlook_mail_entra_audience }}"
+  # #   target: "{{ outlook_mail_entra_mcp_target }}"
+  # - kind: toolbox
+  #   name: agent-tools
+  #   tools:
+  #     - type: web_search
+  #       name: web_search
+  #     - type: code_interpreter
+  #       name: code_interpreter
+  #     # - type: mcp
+  #     #   # This MCP tool doesn't require authentication
+  #     #   server_label: noauth_mcp
+  #     #   server_url: "{{ mcp_endpoint }}"
+  #     #   require_approval: "never"
+  #     - type: mcp
+  #       # This MCP tool uses the GitHub MCP server with a PAT for authentication or OAuth2
+  #       server_label: github
+  #       project_connection_id: github-mcp-pat-conn # use `github-mcp-oauth-conn` for OAuth2 authentication
+  #       require_approval: "never"
+  #     - type: mcp
+  #       # This MCP tool uses the Azure Language MCP server with agent identity for authentication
+  #       server_label: language-mcp
+  #       project_connection_id: language-mcp-conn
+  #       require_approval: "never"
+  #     # - type: mcp
+  #     #   server_label: outlook-mail
+  #     #   project_connection_id: outlook-mail-conn
+  #     #   require_approval: "never"

python/samples/04-hosting/foundry-hosted-agents/responses/04_foundry_toolbox/requirements.txt

@@ -1,3 +1,4 @@
-agent-framework
-agent-framework-foundry-hosting
+# agent-framework
+# agent-framework-foundry-hosting
+
 mcp>=1.24.0,<2

python/scripts/sample_validation/create_dynamic_workflow_executor.py

@@ -14,8 +14,8 @@ from agent_framework import (
     handler,
 )
 from agent_framework.github import GitHubCopilotAgent
-from copilot.session import PermissionHandler, PermissionRequestResult
-from copilot.session_events import PermissionRequest
+from copilot.generated.session_events import PermissionRequest
+from copilot.session import PermissionRequestResult
 from pydantic import BaseModel
 from sample_validation.const import WORKER_COMPLETED
 from sample_validation.discovery import DiscoveryResult
@@ -103,7 +103,7 @@ def prompt_permission(
     logger.debug(
         f"[Permission Request: {request.kind}] ({context})Automatically approved for sample validation."
     )
-    return PermissionHandler.approve_all(request, context)
+    return PermissionRequestResult(kind="approved")
 
 
 class CustomAgentExecutor(Executor):

python/uv.lock

@@ -115,7 +115,7 @@ wheels = [
 
 [[package]]
 name = "agent-framework"
-version = "1.8.0"
+version = "1.7.0"
 source = { virtual = "." }
 dependencies = [
     { name = "agent-framework-core", extra = ["all"], marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -170,7 +170,7 @@ dev = [
 
 [[package]]
 name = "agent-framework-a2a"
-version = "1.0.0b260604"
+version = "1.0.0b260528"
 source = { editable = "packages/a2a" }
 dependencies = [
     { name = "a2a-sdk", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -213,7 +213,7 @@ provides-extras = ["dev"]
 
 [[package]]
 name = "agent-framework-anthropic"
-version = "1.0.0b260604"
+version = "1.0.0b260521"
 source = { editable = "packages/anthropic" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -279,7 +279,7 @@ requires-dist = [
 
 [[package]]
 name = "agent-framework-azurefunctions"
-version = "1.0.0b260604"
+version = "1.0.0b260521"
 source = { editable = "packages/azurefunctions" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -301,7 +301,7 @@ dev = []
 
 [[package]]
 name = "agent-framework-bedrock"
-version = "1.0.0b260604"
+version = "1.0.0b260521"
 source = { editable = "packages/bedrock" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -363,7 +363,7 @@ requires-dist = [
 
 [[package]]
 name = "agent-framework-core"
-version = "1.8.0"
+version = "1.7.0"
 source = { editable = "packages/core" }
 dependencies = [
     { name = "opentelemetry-api", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -529,7 +529,7 @@ dev = [{ name = "types-python-dateutil", specifier = "==2.9.0.20260518" }]
 
 [[package]]
 name = "agent-framework-foundry"
-version = "1.8.0"
+version = "1.7.0"
 source = { editable = "packages/foundry" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -548,7 +548,7 @@ requires-dist = [
 
 [[package]]
 name = "agent-framework-foundry-hosting"
-version = "1.0.0a260604"
+version = "1.0.0a260528"
 source = { editable = "packages/foundry_hosting" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -599,7 +599,7 @@ requires-dist = [
 
 [[package]]
 name = "agent-framework-github-copilot"
-version = "1.0.0rc1"
+version = "1.0.0b260521"
 source = { editable = "packages/github_copilot" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -609,7 +609,7 @@ dependencies = [
 [package.metadata]
 requires-dist = [
     { name = "agent-framework-core", editable = "packages/core" },
-    { name = "github-copilot-sdk", marker = "python_full_version >= '3.11'", specifier = ">=1.0.0,<2" },
+    { name = "github-copilot-sdk", marker = "python_full_version >= '3.11'", specifier = "<=1.0.0b2,>=1.0.0b2" },
 ]
 
 [[package]]
@@ -729,7 +729,7 @@ requires-dist = [
 
 [[package]]
 name = "agent-framework-mistral"
-version = "1.0.0a260604"
+version = "1.0.0a260505"
 source = { editable = "packages/mistral" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -774,7 +774,7 @@ requires-dist = [
 
 [[package]]
 name = "agent-framework-openai"
-version = "1.8.0"
+version = "1.7.0"
 source = { editable = "packages/openai" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -789,7 +789,7 @@ requires-dist = [
 
 [[package]]
 name = "agent-framework-orchestrations"
-version = "1.0.0rc3"
+version = "1.0.0rc2"
 source = { editable = "packages/orchestrations" }
 dependencies = [
     { name = "agent-framework-core", marker = "sys_platform == 'darwin' or sys_platform == 'linux' or sys_platform == 'win32'" },
@@ -2608,19 +2608,19 @@ wheels = [
 
 [[package]]
 name = "github-copilot-sdk"
-version = "1.0.0"
+version = "1.0.0b2"
 source = { registry = "https://pypi.org/simple" }
 dependencies = [
     { name = "pydantic", marker = "(python_full_version >= '3.11' and sys_platform == 'darwin') or (python_full_version >= '3.11' and sys_platform == 'linux') or (python_full_version >= '3.11' and sys_platform == 'win32')" },
     { name = "python-dateutil", marker = "(python_full_version >= '3.11' and sys_platform == 'darwin') or (python_full_version >= '3.11' and sys_platform == 'linux') or (python_full_version >= '3.11' and sys_platform == 'win32')" },
 ]
 wheels = [
-    { url = "https://files.pythonhosted.org/packages/7a/d2/e74fdf476d0dde5c3802b3ba360f1b1e250e55d6d39c03f578c28ac9864e/github_copilot_sdk-1.0.0-py3-none-macosx_10_9_x86_64.whl", hash = "sha256:3cae245fb825e26a74395b74f10d9fd90bc464aa77005848ae0809c9a46c96df", size = 94986104, upload-time = "2026-06-02T14:59:55.022Z" },
-    { url = "https://files.pythonhosted.org/packages/b6/81/e4d9dd01b0a563e488427aa879166287c88de3fccf7b8a95e22a6c652fc3/github_copilot_sdk-1.0.0-py3-none-macosx_11_0_arm64.whl", hash = "sha256:b344a00a877c86ef717244e42bd01acb3694b7377644661c82fc278ccc990e37", size = 91435649, upload-time = "2026-06-02T15:00:02.567Z" },
-    { url = "https://files.pythonhosted.org/packages/bd/ec/e94b8f5a299850e600ffe1fe14bd21b48e01172b9e8b490a0ebd0d0c8d27/github_copilot_sdk-1.0.0-py3-none-manylinux_2_28_aarch64.whl", hash = "sha256:dd3a6b7637a3b12476854aeb599c6bed030f6a166fbd942d872c9a11a695517c", size = 97301959, upload-time = "2026-06-02T15:00:11.019Z" },
-    { url = "https://files.pythonhosted.org/packages/4b/bf/dfba743a11d9745b0664ec5e1ae6e05055a5cbef0ccc6d593222319184eb/github_copilot_sdk-1.0.0-py3-none-manylinux_2_28_x86_64.whl", hash = "sha256:bbd2c64fe37016c74620a02d778eaacbd526b4c3b668a3cdff019f831c752eee", size = 96071193, upload-time = "2026-06-02T15:00:22.634Z" },
-    { url = "https://files.pythonhosted.org/packages/0e/9b/d953dcbb898f4d44efc0cb592e9a703ad43a4b673aafb5bbd763962ab2fd/github_copilot_sdk-1.0.0-py3-none-win_amd64.whl", hash = "sha256:2d46fff634eece978532b1329c0d9e1d784b08ad521e71e6af06c5c28ae2e7c5", size = 90374124, upload-time = "2026-06-02T15:00:31.376Z" },
-    { url = "https://files.pythonhosted.org/packages/f4/f7/0f9943b1439e3dcc52854140676b65d8f63405c471a77c58291a8f4bfb52/github_copilot_sdk-1.0.0-py3-none-win_arm64.whl", hash = "sha256:ebfb80395caa834df8ab16ab4aab3e5d8db883ed3b024f723c394b1514e47221", size = 87874846, upload-time = "2026-06-02T15:00:38.737Z" },
+    { url = "https://files.pythonhosted.org/packages/82/fe/2cb98d4b9f57f8062ea72775bde72aed1958305016753f7296398e0ceb45/github_copilot_sdk-1.0.0b2-py3-none-macosx_10_9_x86_64.whl", hash = "sha256:1b5941d8b6e3d94d42a5bec6607a26f562e6535d5c981089d23d3d224b94601c", size = 67061619, upload-time = "2026-05-06T20:02:08.636Z" },
+    { url = "https://files.pythonhosted.org/packages/57/45/76567821b2d36f81e6bca78c98d265e2762733f765fa51d69602b7f81867/github_copilot_sdk-1.0.0b2-py3-none-macosx_11_0_arm64.whl", hash = "sha256:c5b8f6a087a0cf02bb0d33976e8f8c009578d84d701a0b28d52051304791ac70", size = 63790955, upload-time = "2026-05-06T20:02:12.354Z" },
+    { url = "https://files.pythonhosted.org/packages/15/67/684b0da0b1207a2bdf025c22ee075d34a1736d61a4973651035d4fd4d8dc/github_copilot_sdk-1.0.0b2-py3-none-manylinux_2_28_aarch64.whl", hash = "sha256:f403638c11b82bddb81c94675fc4e8014a1bb2e86a679a39fa167dcc3ad5416a", size = 69538664, upload-time = "2026-05-06T20:02:16.363Z" },
+    { url = "https://files.pythonhosted.org/packages/57/1d/80d88ecf83683535d1a16d4817f1683db3b125f52a924ebdfe9764f5e4c3/github_copilot_sdk-1.0.0b2-py3-none-manylinux_2_28_x86_64.whl", hash = "sha256:433d16bb31171fee8d3a5b70259c527f63b297e83a8f8761ae1f16f14d641f32", size = 68163648, upload-time = "2026-05-06T20:02:21.139Z" },
+    { url = "https://files.pythonhosted.org/packages/32/d3/b72aa2fbb3194b50b53e8cb1484f5606a1f8eedcdb0bfb5747da52079553/github_copilot_sdk-1.0.0b2-py3-none-win_amd64.whl", hash = "sha256:a6e9782dae4c3c2ab3527b45bb5de0f61998104c10e9ff64698280eaf37ab5dd", size = 62649144, upload-time = "2026-05-06T20:02:24.953Z" },
+    { url = "https://files.pythonhosted.org/packages/b6/e2/be95b8ea0ac11d1ca474e28a59284f4e395c2710734eadfb657f5de8ace2/github_copilot_sdk-1.0.0b2-py3-none-win_arm64.whl", hash = "sha256:2e97d0ce4bad67dc5929091cb429e7bbae7d4643e4908a6af256a41439000740", size = 60374365, upload-time = "2026-05-06T20:02:29.02Z" },
 ]
 
 [[package]]