Microsoft Foundry Toolbox

Para um FoundryAgent gerido pelo serviço, anexe a Toolbox à definição do agente no Foundry. As orientações de consumo do .NET Toolbox do lado do cliente não estão atualmente documentadas.

Instale os pacotes

pip install agent-framework-foundry-hosting agent-framework-foundry --pre

FoundryToolbox é importado de agent_framework.foundry e fornecido por agent-framework-foundry-hosting.

Configurar a Caixa de Ferramentas

Defina um endpoint explícito do Toolbox MCP:

TOOLBOX_ENDPOINT="https://<account>.services.ai.azure.com/api/projects/<project>/toolboxes/<name>/mcp?api-version=v1"

Ou deixe que FoundryToolbox crie o endpoint:

FOUNDRY_PROJECT_ENDPOINT="https://<account>.services.ai.azure.com/api/projects/<project>"
TOOLBOX_NAME="<toolbox-name>"

Os exemplos do agente alojado também usam AZURE_AI_MODEL_DEPLOYMENT_NAME para FoundryChatClient.

Utilize FoundryToolbox com um agente alojado

FoundryToolbox resolve o respetivo ponto final, autentica cada pedido MCP com a credencial do Azure fornecida, encaminha o ID de chamada por pedido do Foundry e participa no ciclo de vida da ligação do agente.

import asyncio
import os

from agent_framework import Agent
from agent_framework.foundry import FoundryChatClient, FoundryToolbox, ResponsesHostServer
from azure.identity import DefaultAzureCredential
from dotenv import load_dotenv

# Load environment variables from .env file
load_dotenv()


async def main():
    credential = DefaultAzureCredential()

    # FoundryToolbox resolves the toolbox endpoint from the environment
    # (TOOLBOX_ENDPOINT, or FOUNDRY_PROJECT_ENDPOINT + TOOLBOX_NAME), authenticates
    # every request with the credential, and transparently forwards the platform
    # per-request call-id to the toolbox. The hosting server enters the agent, which
    # connects the toolbox on first use and closes it at shutdown.
    toolbox = FoundryToolbox(credential)

    # Create the chat client
    client = FoundryChatClient(
        project_endpoint=os.environ["FOUNDRY_PROJECT_ENDPOINT"],
        model=os.environ["AZURE_AI_MODEL_DEPLOYMENT_NAME"],
        credential=credential,
    )

    agent = Agent(
        client=client,
        instructions="You are a friendly assistant. Keep your answers brief.",
        tools=toolbox,
        # History will be managed by the hosting infrastructure, thus there
        # is no need to store history by the service. Learn more at:
        # https://developers.openai.com/api/reference/resources/responses/methods/create
        default_options={"store": False},
    )

    server = ResponsesHostServer(agent)
    await server.run_async()

Competências do Expose Toolbox

Uma Caixa de Ferramentas pode expor as Competências do Agente em relação ao MCP. Defina load_tools=False quando apenas as competências devem ser visíveis no modelo, depois adicione a Toolbox como ferramenta para que a sua sessão MCP se conecte e use as_skills_provider() como fornecedora de contexto.

import asyncio
import os

from agent_framework import Agent
from agent_framework.foundry import FoundryChatClient, FoundryToolbox, ResponsesHostServer
from azure.identity import DefaultAzureCredential
from dotenv import load_dotenv

# Load environment variables from .env file
load_dotenv()


async def main() -> None:
    credential = DefaultAzureCredential()

    # FoundryToolbox resolves the toolbox endpoint from the environment
    # (TOOLBOX_ENDPOINT, or FOUNDRY_PROJECT_ENDPOINT + TOOLBOX_NAME), authenticates
    # every request with the credential, and forwards the platform per-request
    # call-id. ``load_tools=False`` keeps the toolbox's tools hidden so only its
    # Agent Skills (SEP-2640) are surfaced; passing it via ``tools=`` connects the
    # MCP session that ``as_skills_provider()`` reads from.
    toolbox = FoundryToolbox(credential, load_tools=False)

    # as_skills_provider() discovers skills from skill://index.json on the toolbox
    # MCP session and exposes them as an agent context provider; SKILL.md bodies are
    # fetched on demand via resources/read. disable_load_skill_approval=True registers
    # the load_skill tool with approval_mode="never_require" so this unattended agent
    # can load skills without an approval round-trip -- the Responses host runs the
    # agent without an AgentSession, which the default approval flow requires.
    skills_provider = toolbox.as_skills_provider(disable_load_skill_approval=True)

    client = FoundryChatClient(
        project_endpoint=os.environ["FOUNDRY_PROJECT_ENDPOINT"],
        model=os.environ["AZURE_AI_MODEL_DEPLOYMENT_NAME"],
        credential=credential,
    )

    agent = Agent(
        client=client,
        name=os.environ.get("AGENT_NAME", "hosted-toolbox-mcp-skills"),
        instructions="You are a helpful assistant.",
        tools=toolbox,
        context_providers=[skills_provider],
        # History will be managed by the hosting infrastructure, thus there
        # is no need to store history by the service. Learn more at:
        # https://developers.openai.com/api/reference/resources/responses/methods/create
        default_options={"store": False},
    )

    server = ResponsesHostServer(agent)
    await server.run_async()

A aprovação continua ativada por predefinição para operações relacionadas com competências. Desative as aprovações individuais apenas para cenários de confiança e não supervisionados.

Utilize uma Caixa de Ferramentas com FoundryAgent

Anexe a caixa de ferramentas à definição de Prompt ou Agente Hospedado no Foundry. FoundryAgent utiliza essa configuração de ferramentas armazenada; passar Toolbox no lado do cliente não o adiciona ao agente gerido.

Liga-te através do MCP bruto

Usa MCPStreamableHTTPTool diretamente quando a aplicação não usa o FoundryToolbox wrapper de alojamento. Forneça o endpoint do Toolbox e um token bearer do Entra ID através de header_provider.

import asyncio
import os
from collections.abc import Callable
from typing import Any, cast

from agent_framework import Agent, MCPStreamableHTTPTool
from agent_framework.foundry import FoundryChatClient
from azure.core.credentials import TokenCredential
from azure.identity import AzureCliCredential, DefaultAzureCredential, get_bearer_token_provider
from dotenv import load_dotenv
def make_toolbox_header_provider(credential: TokenCredential) -> Callable[[dict[str, Any]], dict[str, str]]:
    """Build a header_provider that injects a fresh Azure AI bearer token on every MCP request."""
    get_token = get_bearer_token_provider(credential, "https://ai.azure.com/.default")

    def provide(_kwargs: dict[str, Any]) -> dict[str, str]:
        return {
            "Authorization": f"Bearer {get_token()}",
        }

    return provide


async def main() -> None:
    credential = DefaultAzureCredential()

    toolbox_tool = MCPStreamableHTTPTool(
        name="foundry_toolbox",
        description="Tools exposed by the configured Foundry toolbox",
        url=os.environ["FOUNDRY_TOOLBOX_ENDPOINT"],
        header_provider=make_toolbox_header_provider(credential),
        load_prompts=False,
    )

    async with Agent(
        client=FoundryChatClient(
            project_endpoint=os.environ["FOUNDRY_PROJECT_ENDPOINT"],
            model=os.environ["FOUNDRY_MODEL"],
            credential=credential,
        ),
        instructions="You are a helpful assistant. Use the available toolbox tools to answer the user.",
        tools=toolbox_tool,
    ) as agent:
        query = "What tools do you have access to?"
        print(f"User: {query}")
        result = await agent.run(query)
        print(f"Assistant: {result}")

A amostra de nível mais baixo utiliza FOUNDRY_TOOLBOX_ENDPOINT. A amostra de competências da Toolbox utiliza FOUNDRY_TOOLBOX_MCP_SERVER_URL; estes nomes pertencem a essas amostras e são distintos das definições TOOLBOX_ENDPOINT e TOOLBOX_NAME da classe FoundryToolbox.

Limitações

• As ferramentas MCP dentro de um Toolbox usam autenticação do lado do servidor através de um Foundry project_connection_id; o cliente do Agent Framework não detém o token portador MCP a montante.
• Utilizar a Toolbox como um servidor MCP requer autenticação Entra ID no lado do cliente para o endpoint da Toolbox.
• As respostas de fluxo de consentimento, como a CONSENT_REQUIRED , são tratadas enquanto o agente está em execução, e não enquanto a ligação Toolbox é criada.

Samples

O Go não disponibiliza atualmente um auxiliar do Foundry Toolbox. Configure as Toolboxes no Foundry e utilize declarações de ferramentas locais ou alojadas suportadas para agentes Go.