Uncomment evaluation sample function definitions, keep only invocations commented

[Copilot]

Motivation and Context

Evaluation sample functions were entirely commented out (signatures + bodies), making alternative implementations invisible to users. Only function invocations should be commented to allow easy switching between approaches.

Description

Problem: Function definitions like RunComprehensiveRedTeamEvaluation() and RunWithCustomEvaluator() were wrapped in // comments, hiding their implementations.

Solution: Uncommented all function definitions while keeping invocations commented:

// Main flow - only invocations commented
await RunBasicRedTeamEvaluation(aiProjectClient, agent);  // Active
// await RunComprehensiveRedTeamEvaluation(aiProjectClient, agent);  // Uncomment to use

// Function definitions - all visible
static async Task RunBasicRedTeamEvaluation(...) { /* visible */ }

#pragma warning disable CS8321  // Suppress unused warnings
static async Task RunComprehensiveRedTeamEvaluation(...) { /* visible, ready to use */ }
#pragma warning restore CS8321

Changes:

• Evaluation_Step01_RedTeaming: Uncommented RunComprehensiveRedTeamEvaluation() and RunSimulationOnlyMode()
• Evaluation_Step02_SelfReflection: Uncommented RunWithCustomEvaluator(), EvaluateGroundedness(), and RunWithAzureEvalService()
• Added #pragma warning disable CS8321 around alternative implementations to suppress unused function warnings

Users can now see all available implementations and switch by uncommenting a single invocation line.

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the Contribution Guidelines
• All unit tests pass, and I have added new tests where possible
• Is this a breaking change? If yes, add "[BREAKING]" prefix to the title of the PR.

Original prompt

---

This section details on the original issue you should resolve

<issue_title>.NET: Support for Foundry Observability and Evaluations</issue_title>
<issue_description>Part of #3440

Support for Foundry Observability and Evaluations. Review existing samples and verify everything is working.</issue_description>

<agent_instructions>Follow the instructions in the issue OP and the first comment.</agent_instructions>

Comments on the Issue (you are @copilot in this section)

@rogerbarreto # Subplan: #3675 - Support for Foundry Observability and Evaluations

Issue: #3675
Parent Issue: #3440
Scope: .NET only (parity with Python samples)
Deliverable: Sample files demonstrating Observability and Evaluation features

---

Objective

Create .NET samples demonstrating Foundry Observability and Evaluation capabilities, achieving parity with existing Python samples. This includes verifying existing observability (Step07) works correctly and adding new evaluation samples.

---

Current State Analysis

Existing .NET Samples

Feature	.NET Status	Sample
Observability/Telemetry	✅ Exists	FoundryAgents_Step07_Observability

Missing from .NET (Exists in Python)

Feature	Python Sample	Status
Red Teaming Evaluation	evaluation/red_teaming/red_team_agent_sample.py	❌ Missing
Self Reflection Evaluation	evaluation/self_reflection/self_reflection.py	❌ Missing

---

Workplan

Phase 1: Verify Existing Observability Sample

• Task 1.1: Review existing FoundryAgents_Step07_Observability sample
  • Verify it builds and runs successfully
  • Check telemetry data is being generated
  • Document current capabilities
• Task 1.2: Compare with Python observability patterns
  • Identify any gaps in functionality
  • Check for Azure Monitor integration
• Task 1.3: Update Step07 if needed
  • Ensure parity with Python approach
  • Add any missing telemetry types (if applicable)

Phase 2: Red Teaming Evaluation Sample

• Task 2.1: Analyze Python red teaming sample structure
  • Review red_team_agent_sample.py
  • Review README.md for setup requirements
  • Understand Azure AI Evaluation SDK dependencies
• Task 2.2: Research .NET equivalents
  • Check for Azure.AI.Evaluation NuGet package
  • Identify available .NET evaluation APIs
• Task 2.3: Create evaluation sample directory structure

  dotnet/samples/GettingStarted/Evaluation/
  ├── Evaluation_Step01_RedTeaming/
  │   ├── Evaluation_Step01_RedTeaming.csproj
  │   ├── Program.cs
  │   └── README.md
  

• Task 2.4: Implement Red Teaming sample
  • Configure red team evaluator
  • Set up target callback (FoundryAgent integration)
  • Run evaluation against attack prompts
  • Output evaluation results

Phase 3: Self Reflection Evaluation Sample

• Task 3.1: Analyze Python self-reflection sample
  • Review self_reflection.py
  • Review resources/ folder contents
  • Understand evaluation criteria used
• Task 3.2: Create Self Reflection sample

  dotnet/samples/GettingStarted/Evaluation/
  └── Evaluation_Step02_SelfReflection/
      ├── Evaluation_Step02_SelfReflection.csproj
      ├── Program.cs
      ├── Resources/
      │   └── (evaluation data files)
      └── README.md
  

• Task 3.3: Implement Self Reflection sample
  • Create evaluation agent
  • Configure reflection criteria
  • Run self-assessment
  • Generate quality scores

Phase 4: Documentation & Verification

• Task 4.1: Create Evaluation/README.md
  • Overview of evaluation capabilities
  • Setup instructions (Azure resources needed)
  • Links to Azure AI Evaluation documentation
• Task 4.2: Update main FoundryAgents/README.md
  • Add links to evaluation samples
  • Document observability integration
• Task 4.3: Verify all samples
  • Build verification
  • Run with test data
  • Verify output format

---

Python Sample References

Red Teaming

File	Purpose	GitHub Link
red_team_agent_sample.py	Main sample code	Link
README.md	Setup instructions	Link
.env.example	Environment variables	[Link](https://github.com/microsoft/agent-framework/blob/main/python/samples/getting_started/evaluation/red_teaming/.env...

• Fixes .NET: Support for Foundry Observability and Evaluations #3675

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.