Skip to main content
Many agent frameworks support the ability to request user approval before executing certain actions. This is especially useful when an agent is calling external tools that may have significant effects or costs associated with their usage. The Approval Extension pauses execution and prompts the user with the tool details, resuming only after approval is granted.

​
Basic Implementation

To implement approvals, you inject the extension and call request_approval() within your tool execution logic or a framework handler. This example uses the BeeAI Framework to request user approval before executing a tool call: 
from typing import Annotated, Any

from a2a.types import (
    Message,
)
from agentstack_sdk.server import Server
from agentstack_sdk.server.context import RunContext
from agentstack_sdk.a2a.extensions.interactions.approval import (
    ApprovalExtensionParams,
    ApprovalExtensionServer,
    ApprovalExtensionSpec,
    ToolCallApprovalRequest,
)
from agentstack_sdk.a2a.extensions.tools.exceptions import ToolCallRejectionError
from beeai_framework.agents.requirement import RequirementAgent
from beeai_framework.backend import ChatModel
from beeai_framework.agents.requirement.requirements.ask_permission import AskPermissionRequirement
from beeai_framework.tools import Tool
from beeai_framework.tools.think import ThinkTool
from beeai_framework.adapters.mcp.serve.server import _tool_factory

server = Server()


@server.agent()
async def tool_call_agent(
    input: Message,
    context: RunContext,
    approval_ext: Annotated[ApprovalExtensionServer, ApprovalExtensionSpec(params=ApprovalExtensionParams())],
):
    async def handler(tool: Tool, input: dict[str, Any]) -> bool:

        response = await approval_ext.request_approval(
            # using MCP Tool data model as intermediary to simplify conversion
            ToolCallApprovalRequest.from_mcp_tool(_tool_factory(tool), input=input),  # type: ignore
            context=context,
        )
        return response.approved

    think_tool = ThinkTool()
    agent = RequirementAgent(
        llm=ChatModel.from_name("ollama:gpt-oss:20b"),
        tools=[think_tool],
        requirements=[AskPermissionRequirement([think_tool], handler=handler)],
    )

    result = await agent.run("".join(part.root.text for part in input.parts if part.root.kind == "text"))
    yield result.output[0].text


if __name__ == "__main__":
    server.run()
1

Import the Approval extension

Import ApprovalExtensionServer, ApprovalExtensionSpec, ApprovalExtensionParams, and ToolCallApprovalRequest from agentstack_sdk.a2a.extensions.interactions.approval.
2

Inject the extension

Add an approval parameter to your agent function using the Annotated type hint with ApprovalExtensionServer and ApprovalExtensionSpec.
3

Request user approval

Within your agent logic or tool handler, call await approval_ext.request_approval(). This triggers a UI prompt for the user and pauses execution until they respond.
4

Handle the response

Access the approved boolean from the response. If True, proceed with the tool execution; if False, handle the rejection gracefully (e.g., by yielding a message to the user).