View a markdown version of this page

Buat daftar petunjuk yang tersedia di gateway AgentCore - Batuan Dasar Amazon AgentCore

Buat daftar petunjuk yang tersedia di gateway AgentCore

Untuk mencantumkan semua prompt yang tersedia yang disediakan AgentCore gateway, buat permintaan POST ke titik akhir MCP gateway dan tentukan prompts/list sebagai metode di badan permintaan:

POST /mcp HTTP/1.1 Host: ${GatewayEndpoint} Content-Type: application/json Authorization: ${Authorization header} ${RequestBody}

Ganti nilai-nilai berikut:

catatan

Untuk daftar parameter yang didukung secara opsionalprompts/list, lihat params objek di badan permintaan di Prompts dalam dokumentasi Protokol Konteks Model. Di bagian atas halaman di sebelah bilah pencarian, Anda dapat memilih versi MCP yang dokumentasinya ingin Anda lihat. Pastikan versi tersebut didukung oleh Amazon Bedrock AgentCore.

Respons mengembalikan daftar prompt yang tersedia dengan nama, deskripsi, dan argumennya.

Contoh kode untuk daftar petunjuk

Untuk melihat contoh daftar petunjuk yang tersedia di gateway, pilih salah satu metode berikut:

contoh
Python requests package
  1. import requests import json def list_prompts(gateway_url, access_token): headers = { "Content-Type": "application/json", "Authorization": f"Bearer {access_token}" } payload = { "jsonrpc": "2.0", "id": "list-prompts-request", "method": "prompts/list" } response = requests.post(gateway_url, headers=headers, json=payload) return response.json() # Example usage gateway_url = "https://${GatewayEndpoint}/mcp" # Replace with your actual gateway endpoint access_token = "${AccessToken}" # Replace with your actual access token prompts = list_prompts(gateway_url, access_token) print(json.dumps(prompts, indent=2))
MCP Client
  1. import asyncio from mcp import ClientSession from mcp.client.streamable_http import streamablehttp_client async def execute_mcp( url, token, headers=None ): default_headers = { "Authorization": f"Bearer {token}" } headers = {**default_headers, **(headers or {})} async with streamablehttp_client( url=url, headers=headers, ) as ( read_stream, write_stream, callA, ): async with ClientSession(read_stream, write_stream) as session: # 1. Perform initialization handshake print("Initializing MCP...") _init_response = await session.initialize() print(f"MCP Server Initialize successful! - {_init_response}") # 2. List available prompts print("Listing prompts...") cursor = True prompts = [] while cursor: next_cursor = cursor if isinstance(cursor, bool): next_cursor = None list_prompts_response = await session.list_prompts(next_cursor) prompts.extend(list_prompts_response.prompts) cursor = list_prompts_response.nextCursor prompt_names = [] if prompts: for prompt in prompts: prompt_names.append(prompt.name) prompt_names_string = "\n".join(prompt_names) print( f"List MCP prompts. # of prompts - {len(prompts)}\n" f"List of prompts - \n{prompt_names_string}\n" ) async def main(): url = "https://${GatewayEndpoint}/mcp" token = "your_bearer_token_here" # Optional additional headers additional_headers = { "Content-Type": "application/json", } await execute_mcp( url=url, token=token, headers=additional_headers ) # Run the async function if __name__ == "__main__": asyncio.run(main())
Strands MCP Client
  1. # NOTE: Strands SDK prompt support may vary. Use the MCP Client approach above # for the most reliable prompts/list implementation. from strands.tools.mcp.mcp_client import MCPClient from mcp.client.streamable_http import streamablehttp_client def create_streamable_http_transport(mcp_url: str, access_token: str): return streamablehttp_client(mcp_url, headers={"Authorization": f"Bearer {access_token}"}) def run_agent(mcp_url: str, access_token: str): mcp_client = MCPClient(lambda: create_streamable_http_transport(mcp_url, access_token)) with mcp_client: result = mcp_client.list_prompts_sync() print(f"Found the following prompts: {[prompt.name for prompt in result.prompts]}") run_agent(<MCP URL>, <Access token>)
LangGraph MCP Client
  1. # NOTE: LangGraph MCP adapter prompt support may vary. Use the MCP Client # approach above for the most reliable prompts/list implementation. import asyncio from mcp import ClientSession from mcp.client.streamable_http import streamablehttp_client async def list_prompts(url, token): headers = {"Authorization": f"Bearer {token}"} async with streamablehttp_client(url=url, headers=headers) as ( read_stream, write_stream, callA ): async with ClientSession(read_stream, write_stream) as session: await session.initialize() response = await session.list_prompts() for prompt in response.prompts: print(f"{prompt.name} - {prompt.description}") asyncio.run(list_prompts("https://${GatewayEndpoint}/mcp", "${AccessToken}"))
catatan

Nama prompt diawali dengan nama target menggunakan format{targetName}___{promptName}. Ini adalah konvensi penamaan yang sama yang digunakan untuk alat. Untuk informasi selengkapnya, lihat Konvensi penamaan alat.

catatan

Untuk informasi tentang pembuatan server MCP dengan prompt, lihat fitur stateful MCP.