PrompterOne: Agent for Claude Code

Agent Tools

[!NOTE] The live Learn tools surface now routes through https://learn.microsoft.com/agent-framework/agents/tools/overview. Current live docs also add a provider-support matrix and fold agent-as-tool composition into that broader tools overview.

Tooling support can vary considerably between different agent types. Some agents might allow developers to customize the agent at construction time by providing external function tools or by choosing to activate specific built-in tools that are supported by the agent. On the other hand, some custom agents might support no customization via providing external or activating built-in tools, if they already provide defined features that shouldn't be changed.

::: zone pivot="programming-language-csharp"

Therefore, the base abstraction does not provide any direct tooling support, however each agent can choose whether it accepts tooling customization at construction time.

Tooling support with ChatClientAgent

The ChatClientAgent is an agent class that can be used to build agentic capabilities on top of any inference service. It comes with support for:

1. Using your own function tools with the agent
2. Using built-in tools that the underlying service might support.

[!TIP] For more information on ChatClientAgent and information on supported services, see Simple agents based on inference services

Provide AIFunction instances during agent construction

There are various ways to construct a ChatClientAgent, for example, directly or via factory helper methods on various service clients, but all support passing tools.

// Sample function tool.
[Description("Get the weather for a given location.")]
static string GetWeather([Description("The location to get the weather for.")] string location)
    => $"The weather in {location} is cloudy with a high of 15°C.";

// When calling the ChatClientAgent constructor.
new ChatClientAgent(
    chatClient,
    instructions: "You are a helpful assistant",
    tools: [AIFunctionFactory.Create(GetWeather)]);

// When using one of the helper factory methods.
openAIResponseClient.AsAIAgent(
    instructions: "You are a helpful assistant",
    tools: [AIFunctionFactory.Create(GetWeather)]);