Strands Agents + Amazon Bedrock AgentCoreでA2Aサーバーを構築してデプロイする

{
  "name": "Calculator Agent",
  "description": "A calculator agent that can perform basic arithmetic operations.",
  "url": "http://127.0.0.1:9000/",
  "version": "0.0.1",
  "capabilities": {
    "streaming": true,
    "pushNotifications": false
  },
  "skills": [
    {
      "id": "calculator",
      "name": "calculator",
      "description": "Perform basic arithmetic operations",
      "inputModes": ["text"],
      "outputModes": ["text"]
    }
  ],
  "securitySchemes": {
    "bearer": {
      "type": "http",
      "scheme": "bearer"
    }
  },
  "security": ["bearer"]
}

mkdir a2a-agentcore-demo
cd a2a-agentcore-demo
uv init

uv add 'strands-agents[a2a]' bedrock-agentcore bedrock-agentcore-starter-toolkit strands-agents-tools

import logging
import os
from strands_tools.calculator import calculator
from strands import Agent
from strands.multiagent.a2a import A2AServer
import uvicorn
from fastapi import FastAPI

logging.basicConfig(level=logging.INFO)

# AgentCore Runtimeで動作する場合は環境変数からURLを取得
runtime_url = os.environ.get('AGENTCORE_RUNTIME_URL', 'http://127.0.0.1:9000/')

logging.info(f"Runtime URL: {runtime_url}")

# Strands Agentを作成
strands_agent = Agent(
    name="Calculator Agent",
    description="A calculator agent that can perform basic arithmetic operations.",
    tools=[calculator],
    callback_handler=None
)

host, port = "0.0.0.0", 9000

# A2AServerでラップ
a2a_server = A2AServer(
    agent=strands_agent,
    http_url=runtime_url,
    serve_at_root=True,  # ルートパスでサーブ
    enable_a2a_compliant_streaming=True  # A2A仕様準拠のストリーミング
)

app = FastAPI()

@app.get("/ping")
def ping():
    return {"status": "healthy"}

# A2AサーバーをFastAPIアプリにマウント
app.mount("/", a2a_server.to_fastapi_app())

if __name__ == "__main__":
    uvicorn.run(app, host=host, port=port)

uv run my_a2a_server.py

curl http://localhost:9000/.well-known/agent-card.json | jq .

{
  "capabilities": {
    "streaming": true
  },
  "defaultInputModes": [
    "text"
  ],
  "defaultOutputModes": [
    "text"
  ],
  "description": "A calculator agent that can perform basic arithmetic operations.",
  "name": "Calculator Agent",
  "preferredTransport": "JSONRPC",
  "protocolVersion": "0.3.0",
  "skills": [
    {
      "description": "Calculator powered by SymPy for comprehensive mathematical operations.\n\nThis tool provides advanced mathematical functionality through multiple operation modes,\nincluding expression evaluation, equation solving, calculus operations (derivatives, integrals),\nlimits, series expansions, and matrix operations. Results are formatted with appropriate\nprecision and can be displayed in scientific notation when needed.\n\nHow It Works:\n------------\n1. The function parses the mathematical expression using SymPy's parser\n2. Based on the selected mode, it routes the expression to the appropriate handler\n3. Variables and constants are substituted with their values when provided\n4. The expression is evaluated symbolically and/or numerically as appropriate\n5. Results are formatted based on precision preferences and value magnitude\n6. Rich output is generated with operation details and formatted results\n\nOperation Modes:\n--------------\n- evaluate: Calculate the value of a mathematical expression\n- solve: Find solutions to an equation or system of equations\n- derive: Calculate derivatives of an expression\n- integrate: Find the indefinite integral of an expression\n- limit: Evaluate the limit of an expression at a point\n- series: Generate series expansion of an expression\n- matrix: Perform matrix operations\n\nCommon Usage Scenarios:\n---------------------\n- Basic calculations: Evaluating arithmetic expressions\n- Equation solving: Finding roots of polynomials or systems of equations\n- Calculus: Computing derivatives and integrals for analysis\n- Engineering analysis: Working with scientific notations and constants\n- Mathematics education: Visualizing step-by-step solutions\n- Data science: Matrix operations and statistical calculations\n\nReturns:\n    Dict containing status and response content in the format:\n    {\n        \"status\": \"success|error\",\n        \"content\": [{\"text\": \"Result: <calculated_result>\"}]\n    }\n\n    Success case: Returns the calculation result with appropriate formatting\n    Error case: Returns information about what went wrong during calculation\n\nNotes:\n    - For equation solving, set the expression equal to zero implicitly (x**2 + 1 means x**2 + 1 = 0)\n    - Use 'pi' and 'e' for mathematical constants\n    - The 'wrt' parameter is required for differentiation and integration\n    - Matrix expressions use Python-like syntax: [[1, 2], [3, 4]]\n    - Precision control impacts display only, internal calculations use higher precision\n    - Symbolic results are returned when possible unless force_numeric=True",
      "id": "calculator",
      "name": "calculator",
      "tags": []
    }
  ],
  "url": "http://127.0.0.1:9000/",
  "version": "0.0.1"
}

curl -X POST http://localhost:9000 \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": "req-001",
    "method": "message/send",
    "params": {
      "message": {
        "role": "user",
        "parts": [
          {
            "kind": "text",
            "text": "what is 101 * 11?"
          }
        ],
        "messageId": "12345678-1234-1234-1234-123456789012"
      }
    }
  }' | jq .

{
  "id": "req-001",
  "jsonrpc": "2.0",
  "result": {
    "artifacts": [
      {
        "artifactId": "f5d3aef1-f914-40c5-a6db-caff80ca4e6c",
        "name": "agent_response",
        "parts": [
          {
            "kind": "text",
            "text": "101 * 11 = 1111\n"
          }
        ]
      }
    ],
    "contextId": "584a8626-1cc5-45ac-9e87-22ea758c007f",
    "history": [
      {
        "contextId": "584a8626-1cc5-45ac-9e87-22ea758c007f",
        "kind": "message",
        "messageId": "12345678-1234-1234-1234-123456789012",
        "parts": [
          {
            "kind": "text",
            "text": "what is 101 * 11?"
          }
        ],
        "role": "user",
        "taskId": "75018576-695b-4e86-919a-2db4f277c23a"
      },
      {
        "contextId": "584a8626-1cc5-45ac-9e87-22ea758c007f",
        "kind": "message",
        "messageId": "ad5f78f5-b9ba-4f6e-b57e-3a10e80ef930",
        "parts": [
          {
            "kind": "text",
            "text": "101 * 11 = "
          }
        ],
        "role": "agent",
        "taskId": "75018576-695b-4e86-919a-2db4f277c23a"
      },
      {
        "contextId": "584a8626-1cc5-45ac-9e87-22ea758c007f",
        "kind": "message",
        "messageId": "417d0a84-00c6-41d1-b59d-f1b85c384c3c",
        "parts": [
          {
            "kind": "text",
            "text": "1111"
          }
        ],
        "role": "agent",
        "taskId": "75018576-695b-4e86-919a-2db4f277c23a"
      }
    ],
    "id": "75018576-695b-4e86-919a-2db4f277c23a",
    "kind": "task",
    "status": {
      "state": "completed",
      "timestamp": "2026-01-28T12:49:50.490302+00:00"
    }
  }
}

uv run agentcore configure -e my_a2a_server.py --protocol A2A

uv run agentcore deploy

Agent Runtime ARN: arn:aws:bedrock-agentcore:us-west-2:123456789012:runtime/my_a2a_server-xyz123

uv run agentcore invoke '{"jsonrpc": "2.0", "id": "req-001", "method": "message/send", "params": {"message": {"role": "user", "parts": [{"kind": "text", "text": "what is 101 * 11?"}], "messageId": "12345678-1234-1234-1234-123456789012"}}}'

{
  "id": "req-001",
  "jsonrpc": "2.0",
  "result": {
    "artifacts": [
      {
        "artifactId": "19989fb9-05e9-4b3e-9342-8bfd6aaf9020",
        "name": "agent_response",
        "parts": [
          {"kind": "text", "text": "I'll calculate 101 * "},
          {"kind": "text", "text": "11 for you."},
          {"kind": "text", "text": "101"},
          {"kind": "text", "text": " * 11 "},
          {"kind": "text", "text": "= 1111"}
        ]
      }
    ],
    "status": {"state": "completed"}
  }
}

# デプロイしたエージェントを削除
uv run agentcore destroy

#!/bin/bash

REGION="us-west-2"

# User Pool作成
export POOL_ID=$(aws cognito-idp create-user-pool \
  --pool-name "A2AAgentUserPool" \
  --policies '{"PasswordPolicy":{"MinimumLength":8}}' \
  --region $REGION | jq -r '.UserPool.Id')

# App Client作成（クライアントシークレットなし）
export CLIENT_ID=$(aws cognito-idp create-user-pool-client \
  --user-pool-id $POOL_ID \
  --client-name "A2AAgentClient" \
  --no-generate-secret \
  --explicit-auth-flows "ALLOW_USER_PASSWORD_AUTH" "ALLOW_REFRESH_TOKEN_AUTH" \
  --region $REGION | jq -r '.UserPoolClient.ClientId')

# テストユーザー作成
aws cognito-idp admin-create-user \
  --user-pool-id $POOL_ID \
  --username "testuser" \
  --region $REGION \
  --message-action SUPPRESS > /dev/null

# パスワード設定（本番環境では安全なパスワードを使用してください）
aws cognito-idp admin-set-user-password \
  --user-pool-id $POOL_ID \
  --username "testuser" \
  --password "TestPassword123!" \
  --region $REGION \
  --permanent > /dev/null

# Bearerトークン取得
export BEARER_TOKEN=$(aws cognito-idp initiate-auth \
  --client-id "$CLIENT_ID" \
  --auth-flow USER_PASSWORD_AUTH \
  --auth-parameters USERNAME='testuser',PASSWORD='TestPassword123!' \
  --region $REGION | jq -r '.AuthenticationResult.AccessToken')

# Discovery URL
export DISCOVERY_URL="https://cognito-idp.$REGION.amazonaws.com/$POOL_ID/.well-known/openid-configuration"

# 結果出力
echo "=========================================="
echo "Pool ID: $POOL_ID"
echo "Client ID: $CLIENT_ID"
echo "Discovery URL: $DISCOVERY_URL"
echo "Bearer Token: $BEARER_TOKEN"
echo "=========================================="

source setup_cognito.sh

uv run agentcore configure -e my_a2a_server.py --name my_a2a_server_oauth --protocol A2A

uv run agentcore deploy

uv run agentcore invoke '{"jsonrpc": "2.0", "id": "req-001", "method": "message/send", "params": {"message": {"role": "user", "parts": [{"kind": "text", "text": "what is 101 * 11?"}], "messageId": "12345678-1234-1234-1234-123456789012"}}}' --bearer-token $BEARER_TOKEN

Response:
{"id":"req-001","jsonrpc":"2.0","result":{"artifacts":[{"artifactId":"362349e7-309b-4b28-836c-3fe0f669225e","name":"agent_response","parts":[{"kind
":"text","text":"I"},{"kind":"text","text":"'ll calculate 101 "},{"kind":"text","text":"* 11 for"},{"kind":"text","text":"
you."},{"kind":"text","text":"The result"},{"kind":"text","text":" of"},{"kind":"text","text":" 101 * "},{"kind":"text","text":"11 is
**1"},{"kind":"text","text":"111**."}]}],"contextId":"7c2f7cce-0af6-4ea9-b85a-fc57c47033db","history":[{"contextId":"7c2f7cce-0af6-4ea9-b85a-fc57c4
7033db","kind":"message","messageId":"12345678-1234-1234-1234-123456789012","parts":[{"kind":"text","text":"what is 101 *
11?"}],"role":"user","taskId":"6ff8b0c0-ed8d-4084-bff6-3dcc38adbd16"}],"id":"6ff8b0c0-ed8d-4084-bff6-3dcc38adbd16","kind":"task","status":{"state":
"completed","timestamp":"2026-02-01T00:49:17.845839+00:00"}}}

import asyncio
import os
from uuid import uuid4
import httpx
from a2a.client import A2ACardResolver, ClientConfig, ClientFactory
from a2a.types import Message, Part, Role, TextPart

async def send_message(message: str):
    # 環境変数から設定を取得
    runtime_url = os.environ.get('AGENTCORE_RUNTIME_URL')
    bearer_token = os.environ.get('BEARER_TOKEN')
    session_id = str(uuid4())

    # JWT認証ヘッダーを設定
    headers = {
        "Authorization": f"Bearer {bearer_token}",
        "X-Amzn-Bedrock-AgentCore-Runtime-Session-Id": session_id
    }

    async with httpx.AsyncClient(timeout=300, headers=headers) as httpx_client:
        # Agent Cardを取得
        resolver = A2ACardResolver(httpx_client=httpx_client, base_url=runtime_url)
        agent_card = await resolver.get_agent_card()
        print(f"Agent: {agent_card.name}")
        print(f"Skills: {[s.name for s in agent_card.skills]}")

        # クライアントを作成してメッセージ送信
        config = ClientConfig(httpx_client=httpx_client, streaming=False)
        factory = ClientFactory(config)
        client = factory.create(agent_card)

        msg = Message(
            kind="message",
            role=Role.user,
            parts=[Part(TextPart(kind="text", text=message))],
            message_id=uuid4().hex,
        )

        async for event in client.send_message(msg):
            if isinstance(event, Message):
                print(f"Response: {event.model_dump_json(exclude_none=True, indent=2)}")
                return event

if __name__ == "__main__":
    asyncio.run(send_message("what is 101 * 11?"))

# 元のARN
arn:aws:bedrock-agentcore:us-west-2:123456789012:runtime/my_a2a_server_oauth-xxxxxxxx

# URLエンコード後
arn%3Aaws%3Abedrock-agentcore%3Aus-west-2%3A123456789012%3Aruntime%2Fmy_a2a_server_oauth-xxxxxxxx

from urllib.parse import quote

arn = "arn:aws:bedrock-agentcore:us-west-2:123456789012:runtime/my_a2a_server_oauth-xxxxxxxx"
encoded_arn = quote(arn, safe='')
print(encoded_arn)
# 出力: arn%3Aaws%3Abedrock-agentcore%3Aus-west-2%3A123456789012%3Aruntime%2Fmy_a2a_server_oauth-xxxxxxxx

export AGENTCORE_RUNTIME_URL="https://bedrock-agentcore.us-west-2.amazonaws.com/runtimes/arn%3Aaws%3Abedrock-agentcore%3Aus-west-2%3A123456789012%3Aruntime%2Fmy_a2a_server_oauth-xxxxxxxx/invocations/"
export BEARER_TOKEN="<Bearer Token>"
uv run a2a_oauth_client.py

Agent: Calculator Agent
Skills: ['calculator']

Sending message: what is 101 * 11?

Task status: completed
Response: I'll calculate 101 * 11 for you.The result of 101 * 11 is **1111**.

uv add 'strands-agents-tools[a2a_client]'

import os
from strands import Agent
from strands_tools.a2a_client import A2AClientToolProvider

# 環境変数から設定を取得
runtime_url = os.environ.get('AGENTCORE_RUNTIME_URL')
bearer_token = os.environ.get('BEARER_TOKEN')

# JWT認証ヘッダーを設定
httpx_client_args = {
    "headers": {
        "Authorization": f"Bearer {bearer_token}",
    },
    "timeout": 300
}

# A2AClientToolProviderを作成
# known_agent_urls にリモートエージェントのURLを指定
provider = A2AClientToolProvider(
    known_agent_urls=[runtime_url],
    httpx_client_args=httpx_client_args
)

# オーケストレーターエージェントを作成
# A2AClientToolProviderが提供するツールを渡す
orchestrator = Agent(
    name="Orchestrator Agent",
    description="リモートのA2Aエージェントと連携するオーケストレーター",
    tools=provider.tools
)

# オーケストレーターに指示を出す
# 内部でA2Aサーバーを呼び出して計算してくれる
response = orchestrator("リモートの計算エージェントを使って、256 * 16 を計算してください")

uv run orchestrator_agent.py

リモートの計算エージェントを使って256 × 16を計算いたします。まず、利用可能なエージェントを確認してから計算を実行します。
Tool #1: a2a_list_discovered_agents
計算エージェントが利用可能です。それでは256 × 16を計算してもらいます。
Tool #2: a2a_send_message
リモートの計算エージェントを使って計算した結果、**256 × 16 = 4096** です。

計算エージェントが正確に計算を実行し、結果を返してくれました。リモートの計算エージェントを使って計算した結果、**256 × 16 = 4096** です。

uv run agentcore destroy

aws cognito-idp delete-user-pool --user-pool-id $POOL_ID --region us-west-2