MCP com FastMCP (3/4): resources avançados e prompts

MCP com FastMCP (3/4): resources avançados e prompts

Nas partes anteriores criamos um servidor MCP com tools (Parte 1) e vimos a camada de transporte, o contexto e os primeiros resources (Parte 2). Agora damos mais um passo: adicionamos **contexto aos resources**, criamos **resource templates** parametrizados e definimos **prompts** reutilizáveis.

⚠️ Este capítulo continua o código das partes anteriores. Certifique-se de ter o ambiente e o servidor das partes 1 e 2.

Aviso: Este post foi traduzido para o português usando um modelo de tradução automática. Por favor, me avise se encontrar algum erro.

📚 **Esta entrada faz parte da série _MCP com FastMCP_**, dividida em quatro capítulos que são lidos em ordem:

> * Parte 1: Primeiro servidor e ferramentas

* Parte 2: Transporte, contexto y resources

* 👉 **Parte 3: Recursos avançados e prompts**

* Parte 4: HTTP, autenticação e cliente

Adicionar contexto ao resourcelink image 25

Assim como fizemos com as tools, podemos adicionar contexto aos resources.

Servidor MCPlink image 26

	
< > Input
Python
%%writefile gitHub_MCP_server/github_server.py
import httpx
from fastmcp import FastMCP, Context
from github import GITHUB_TOKEN, create_github_headers
USER_ID = 1234567890
# Create FastMCP server
mcp = FastMCP(
name="GitHubMCP",
instructions="This server provides tools, resources and prompts to interact with the GitHub API.",
include_tags={"public"},
exclude_tags={"first_issue"}
)
sub_mcp = FastMCP(
name="SubMCP",
)
@mcp.tool(
tags={"public", "production"},
exclude_args=["user_id"], # user_id has to be injected by server, not provided by LLM
)
async def list_repository_issues(owner: str, repo_name: str, ctx: Context, user_id: int = USER_ID) -&gt; list[dict]:
"""
Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
ctx: The context of the request
user_id: The user ID (automatically injected by the server)
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
"""
# Limit to first 10 issues to avoid very long responses
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=open&amp;per_page=10"
ctx.info(f"Fetching issues from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
ctx.info("No open issues found for this repository.")
return [{"message": "No open issues found for this repository."}]
issues_summary = []
for issue in issues_data:
# Create a more concise summary
summary = f"#{issue.get('number', 'N/A')}: {issue.get('title', 'No title')}"
if issue.get('comments', 0) &gt; 0:
summary += f" ({issue.get('comments')} comments)"
issues_summary.append({
"number": issue.get("number"),
"title": issue.get("title"),
"user": issue.get("user", {}).get("login"),
"url": issue.get("html_url"),
"comments": issue.get("comments"),
"summary": summary
})
ctx.info(f"Found {len(issues_summary)} open issues.")
# Add context information
result = {
"total_found": len(issues_summary),
"repository": f"{owner}/{repo_name}",
"note": "Showing first 10 open issues" if len(issues_summary) == 10 else f"Showing all {len(issues_summary)} open issues",
"issues": issues_summary,
"requested_by_user_id": user_id
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
ctx.error(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
ctx.error(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.tool(tags={"private", "development"})
async def list_more_repository_issues(owner: str, repo_name: str) -&gt; list[dict]:
"""
Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
"""
# Limit to first 100 issues to avoid very long responses
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=open&amp;per_page=100"
print(f"Fetching issues from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
print("No open issues found for this repository.")
return [{"message": "No open issues found for this repository."}]
issues_summary = []
for issue in issues_data:
# Create a more concise summary
summary = f"#{issue.get('number', 'N/A')}: {issue.get('title', 'No title')}"
if issue.get('comments', 0) &gt; 0:
summary += f" ({issue.get('comments')} comments)"
issues_summary.append({
"number": issue.get("number"),
"title": issue.get("title"),
"user": issue.get("user", {}).get("login"),
"url": issue.get("html_url"),
"comments": issue.get("comments"),
"summary": summary
})
print(f"Found {len(issues_summary)} open issues.")
# Add context information
result = {
"total_found": len(issues_summary),
"repository": f"{owner}/{repo_name}",
"note": "Showing first 10 open issues" if len(issues_summary) == 10 else f"Showing all {len(issues_summary)} open issues",
"issues": issues_summary
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
print(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
print(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.tool(tags={"public", "first_issue"})
async def first_repository_issue(owner: str, repo_name: str) -&gt; list[dict]:
"""
Gets the first issue ever created in a GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
Returns:
list[dict]: A list containing information about the first issue created
"""
# Get the first issue by sorting by creation date in ascending order
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=all&amp;sort=created&amp;direction=asc&amp;per_page=1"
print(f"Fetching first issue from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
print("No issues found for this repository.")
return [{"message": "No issues found for this repository."}]
first_issue = issues_data[0]
# Create a detailed summary of the first issue
summary = f"#{first_issue.get('number', 'N/A')}: {first_issue.get('title', 'No title')}"
if first_issue.get('comments', 0) &gt; 0:
summary += f" ({first_issue.get('comments')} comments)"
issue_info = {
"number": first_issue.get("number"),
"title": first_issue.get("title"),
"user": first_issue.get("user", {}).get("login"),
"url": first_issue.get("html_url"),
"state": first_issue.get("state"),
"comments": first_issue.get("comments"),
"created_at": first_issue.get("created_at"),
"updated_at": first_issue.get("updated_at"),
"body": first_issue.get("body", "")[:500] + "..." if len(first_issue.get("body", "")) &gt; 500 else first_issue.get("body", ""),
"summary": summary
}
print(f"Found first issue: #{first_issue.get('number')} created on {first_issue.get('created_at')}")
# Add context information
result = {
"repository": f"{owner}/{repo_name}",
"note": "This is the very first issue created in this repository",
"first_issue": issue_info
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
print(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
print(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.resource("resource://server_info", tags={"public"})
def server_info(ctx: Context) -&gt; str:
"""
Returns information about the server.
"""
return {
"info": "This is the MCP GitHub server development for MaximoFN blog post",
"requested_id": ctx.request_id
}
@sub_mcp.tool(tags={"public"})
def hello_world() -&gt; str:
"""
Returns a simple greeting.
"""
return "Hello, world!"
mcp.mount("sub_mcp", sub_mcp)
if __name__ == "__main__":
print("DEBUG: Starting FastMCP GitHub server...")
print(f"DEBUG: Server name: {mcp.name}")
# Initialize and run the server
mcp.run(
transport="stdio"
)
Copied
>_ Output
			
Overwriting gitHub_MCP_server/github_server.py

Adicionamos contexto ao resource server_info para que ele nos retorne a ID da solicitação.

return {
"info": "Esta es la publicación del blog sobre el desarrollo del servidor MCP GitHub para MaximoFN",
"requested_id": ctx.request_id
}

Teste do servidor com contexto no resourcelink image 27

Executamos o cliente

	
< > Input
Python
!cd client_MCP && source .venv/bin/activate && python client.py ../gitHub_MCP_server/github_server.py
Copied
>_ Output
			
🔗 Connecting to FastMCP server: ../gitHub_MCP_server/github_server.py
✅ Client created successfully
/Users/macm1/Documents/web/portafolio/posts/gitHub_MCP_server/github_server.py:258: DeprecationWarning: Mount prefixes are now optional and the first positional argument should be the server you want to mount.
mcp.mount("sub_mcp", sub_mcp)
[06/28/25 11:17:41] INFO Starting MCP server 'GitHubMCP' with transport 'stdio' server.py:1246
🛠️ Available tools (2):
==================================================
📋 sub_mcp_hello_world
Description: Returns a simple greeting.
Parameters:
📋 list_repository_issues
Description: Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
ctx: The context of the request
user_id: The user ID (automatically injected by the server)
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
Parameters: owner, repo_name
...
🔧 Claude wants to use: read_mcp_resource
📝 Arguments: {'resource_uri': 'resource://server_info'}
📖 Resource read successfully: resource://server_info
🤖 Claude: I'll help you read the server information using the `read_mcp_resource` function. The server information is available at the resource URI "resource://server_info".According to the server information:
- This is the MCP GitHub server development for MaximoFN blog post
- The requested ID is "7"
👤 You: q
👋 Bye!
🧹 Cleaning up resources...
✅ Resources released

Como vemos, ele nos deu as informações do servidor e o ID da solicitação.

👤 Você: Me diga as informações do servidor

🤔 Claude está pensando...
🔧 Claude quiere usar: read_mcp_resource
📝 Argumentos: {'resource_uri': 'resource://server_info'}
📖 Recurso lido com sucesso: resource://server_info
🤖 Claude: Vou ajudá-lo a ler as informações do servidor usando a função `read_mcp_resource`. As informações do servidor estão disponíveis no URI do recurso "resource://server_info".De acordo com as informações do servidor:
- Este es el post del blog de desarrollo del servidor MCP de GitHub para MaximoFN
- O ID solicitado é "7"

Criar um resource templatelink image 28

Anteriormente, criamos um resource que é um recurso estático, mas talvez queiramos obter informação, porém não sempre a mesma; queremos que o LLM possa decidir qual informação quer ou precisa.

Para isso, temos os resource templates, que nos fornecem informações iguais às de um resource, mas de forma dinâmica. No momento da solicitação, o resource é criado e retornado.

Servidor MCPlink image 29

Criar um resource template é feito da mesma maneira que criar um resource, ou seja, por meio de @mcp.resource(<ENDPOINT), só que agora o endpoint é um modelo que é preenchido no momento da solicitação.

Vamos ver isso

	
< > Input
Python
%%writefile gitHub_MCP_server/github_server.py
import httpx
from fastmcp import FastMCP, Context
from github import GITHUB_TOKEN, create_github_headers
import datetime
USER_ID = 1234567890
# Create FastMCP server
mcp = FastMCP(
name="GitHubMCP",
instructions="This server provides tools, resources and prompts to interact with the GitHub API.",
include_tags={"public"},
exclude_tags={"first_issue"}
)
sub_mcp = FastMCP(
name="SubMCP",
)
@mcp.tool(
tags={"public", "production"},
exclude_args=["user_id"], # user_id has to be injected by server, not provided by LLM
)
async def list_repository_issues(owner: str, repo_name: str, ctx: Context, user_id: int = USER_ID) -&gt; list[dict]:
"""
Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
ctx: The context of the request
user_id: The user ID (automatically injected by the server)
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
"""
# Limit to first 10 issues to avoid very long responses
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=open&amp;per_page=10"
ctx.info(f"Fetching issues from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
ctx.info("No open issues found for this repository.")
return [{"message": "No open issues found for this repository."}]
issues_summary = []
for issue in issues_data:
# Create a more concise summary
summary = f"#{issue.get('number', 'N/A')}: {issue.get('title', 'No title')}"
if issue.get('comments', 0) &gt; 0:
summary += f" ({issue.get('comments')} comments)"
issues_summary.append({
"number": issue.get("number"),
"title": issue.get("title"),
"user": issue.get("user", {}).get("login"),
"url": issue.get("html_url"),
"comments": issue.get("comments"),
"summary": summary
})
ctx.info(f"Found {len(issues_summary)} open issues.")
# Add context information
result = {
"total_found": len(issues_summary),
"repository": f"{owner}/{repo_name}",
"note": "Showing first 10 open issues" if len(issues_summary) == 10 else f"Showing all {len(issues_summary)} open issues",
"issues": issues_summary,
"requested_by_user_id": user_id
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
ctx.error(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
ctx.error(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.tool(tags={"private", "development"})
async def list_more_repository_issues(owner: str, repo_name: str) -&gt; list[dict]:
"""
Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
"""
# Limit to first 100 issues to avoid very long responses
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=open&amp;per_page=100"
print(f"Fetching issues from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
print("No open issues found for this repository.")
return [{"message": "No open issues found for this repository."}]
issues_summary = []
for issue in issues_data:
# Create a more concise summary
summary = f"#{issue.get('number', 'N/A')}: {issue.get('title', 'No title')}"
if issue.get('comments', 0) &gt; 0:
summary += f" ({issue.get('comments')} comments)"
issues_summary.append({
"number": issue.get("number"),
"title": issue.get("title"),
"user": issue.get("user", {}).get("login"),
"url": issue.get("html_url"),
"comments": issue.get("comments"),
"summary": summary
})
print(f"Found {len(issues_summary)} open issues.")
# Add context information
result = {
"total_found": len(issues_summary),
"repository": f"{owner}/{repo_name}",
"note": "Showing first 10 open issues" if len(issues_summary) == 10 else f"Showing all {len(issues_summary)} open issues",
"issues": issues_summary
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
print(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
print(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.tool(tags={"public", "first_issue"})
async def first_repository_issue(owner: str, repo_name: str) -&gt; list[dict]:
"""
Gets the first issue ever created in a GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
Returns:
list[dict]: A list containing information about the first issue created
"""
# Get the first issue by sorting by creation date in ascending order
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=all&amp;sort=created&amp;direction=asc&amp;per_page=1"
print(f"Fetching first issue from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
print("No issues found for this repository.")
return [{"message": "No issues found for this repository."}]
first_issue = issues_data[0]
# Create a detailed summary of the first issue
summary = f"#{first_issue.get('number', 'N/A')}: {first_issue.get('title', 'No title')}"
if first_issue.get('comments', 0) &gt; 0:
summary += f" ({first_issue.get('comments')} comments)"
issue_info = {
"number": first_issue.get("number"),
"title": first_issue.get("title"),
"user": first_issue.get("user", {}).get("login"),
"url": first_issue.get("html_url"),
"state": first_issue.get("state"),
"comments": first_issue.get("comments"),
"created_at": first_issue.get("created_at"),
"updated_at": first_issue.get("updated_at"),
"body": first_issue.get("body", "")[:500] + "..." if len(first_issue.get("body", "")) &gt; 500 else first_issue.get("body", ""),
"summary": summary
}
print(f"Found first issue: #{first_issue.get('number')} created on {first_issue.get('created_at')}")
# Add context information
result = {
"repository": f"{owner}/{repo_name}",
"note": "This is the very first issue created in this repository",
"first_issue": issue_info
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
print(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
print(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.resource("resource://server_info", tags={"public"})
def server_info(ctx: Context) -&gt; str:
"""
Returns information about the server.
"""
return {
"info": "This is the MCP GitHub server development for MaximoFN blog post",
"requested_id": ctx.request_id
}
@mcp.resource("github://repo/{owner}/{repo_name}", tags={"public"})
async def repository_info(owner: str, repo_name: str, ctx: Context) -&gt; dict:
"""
Returns detailed information about a GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
ctx: The context of the request
Returns:
dict: Repository information including name, description, stats, etc.
"""
api_url = f"https://api.github.com/repos/{owner}/{repo_name}"
ctx.info(f"Fetching repository information from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
repo_data = response.json()
# Extract relevant repository information
repo_info = {
"name": repo_data.get("name"),
"full_name": repo_data.get("full_name"),
"description": repo_data.get("description"),
"owner": {
"login": repo_data.get("owner", {}).get("login"),
"type": repo_data.get("owner", {}).get("type")
},
"html_url": repo_data.get("html_url"),
"clone_url": repo_data.get("clone_url"),
"ssh_url": repo_data.get("ssh_url"),
"language": repo_data.get("language"),
"size": repo_data.get("size"), # Size in KB
"stargazers_count": repo_data.get("stargazers_count"),
"watchers_count": repo_data.get("watchers_count"),
"forks_count": repo_data.get("forks_count"),
"open_issues_count": repo_data.get("open_issues_count"),
"default_branch": repo_data.get("default_branch"),
"created_at": repo_data.get("created_at"),
"updated_at": repo_data.get("updated_at"),
"pushed_at": repo_data.get("pushed_at"),
"is_private": repo_data.get("private"),
"is_fork": repo_data.get("fork"),
"is_archived": repo_data.get("archived"),
"has_issues": repo_data.get("has_issues"),
"has_projects": repo_data.get("has_projects"),
"has_wiki": repo_data.get("has_wiki"),
"license": repo_data.get("license", {}).get("name") if repo_data.get("license") else None,
"topics": repo_data.get("topics", [])
}
ctx.info(f"Successfully retrieved information for repository {owner}/{repo_name}")
return repo_info
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 404:
error_message = f"Repository {owner}/{repo_name} not found or is private."
elif e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
ctx.error(f"GitHub API error: {e.response.status_code}. {error_message}")
return {
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message,
"repository": f"{owner}/{repo_name}"
}
except Exception as e:
ctx.error(f"An unexpected error occurred: {str(e)}")
return {
"error": f"An unexpected error occurred: {str(e)}",
"repository": f"{owner}/{repo_name}"
}
@sub_mcp.tool(tags={"public"})
def hello_world() -&gt; str:
"""
Returns a simple greeting.
"""
return "Hello, world!"
mcp.mount("sub_mcp", sub_mcp)
if __name__ == "__main__":
print("DEBUG: Starting FastMCP GitHub server...")
print(f"DEBUG: Server name: {mcp.name}")
# Initialize and run the server
mcp.run(
transport="stdio"
)
Copied
>_ Output
			
Overwriting gitHub_MCP_server/github_server.py

Criámos o resource template repository_info, que nos fornece a informação de um repositório que vai determinar o LLM. A template é criada e, em tempo de execução, é preenchida com os parâmetros que lhe são passados.

@mcp.resource("github://repo/{owner}/{repo_name}", tags={"public"})

python

async def repository_info(owner: str, repo_name: str, ctx: Context) -> dict:

Tanto o repositório quanto o proprietário do repositório têm que ser parâmetros da função.

Cliente MCPlink image 30

Fazemos uma pequena alteração no cliente para que o LLM entenda que existem recursos estáticos e dinâmicos

	
< > Input
Python
%%writefile client_MCP/client.py
import sys
import asyncio
from contextlib import AsyncExitStack
from anthropic import Anthropic
from dotenv import load_dotenv
from fastmcp import Client
# Load environment variables from .env file
load_dotenv()
class FastMCPClient:
"""
FastMCP client that integrates with Claude to process user queries
and use tools and resources exposed by a FastMCP server.
"""
def __init__(self):
"""Initialize the FastMCP client with Anthropic and resource management."""
self.exit_stack = AsyncExitStack()
self.anthropic = Anthropic()
self.client = None
async def connect_to_server(self, server_script_path: str):
"""
Connect to the specified FastMCP server.
Args:
server_script_path: Path to the server script (Python)
"""
print(f"🔗 Connecting to FastMCP server: {server_script_path}")
# Determine the server type based on the extension
if not server_script_path.endswith('.py'):
raise ValueError(f"Unsupported server type. Use .py files. Received: {server_script_path}")
# Create FastMCP client
self.client = Client(server_script_path)
# Note: FastMCP Client automatically infers transport from .py files
print("✅ Client created successfully")
async def list_available_tools(self):
"""List available tools in the FastMCP server."""
try:
# Get list of tools from the server using FastMCP context
async with self.client as client:
tools = await client.list_tools()
if tools:
print(f" 🛠️ Available tools ({len(tools)}):")
print("=" * 50)
for tool in tools:
print(f"📋 {tool.name}")
if tool.description:
print(f" Description: {tool.description}")
# Show parameters if available
if hasattr(tool, 'inputSchema') and tool.inputSchema:
if 'properties' in tool.inputSchema:
params = list(tool.inputSchema['properties'].keys())
print(f" Parameters: {', '.join(params)}")
print()
else:
print("⚠️ No tools found in the server")
except Exception as e:
print(f"❌ Error listing tools: {str(e)}")
async def list_available_resources(self):
"""List available resources in the FastMCP server."""
try:
# Get list of resources from the server using FastMCP context
async with self.client as client:
resources = await client.list_resources()
if resources:
print(f" 📚 Available resources ({len(resources)}):")
print("=" * 50)
for resource in resources:
print(f"📄 {resource.uri}")
if resource.name:
print(f" Name: {resource.name}")
if resource.description:
print(f" Description: {resource.description}")
if resource.mimeType:
print(f" MIME Type: {resource.mimeType}")
print()
else:
print("⚠️ No resources found in the server")
except Exception as e:
print(f"❌ Error listing resources: {str(e)}")
async def read_resource(self, resource_uri: str):
"""
Read a specific resource from the server.
Args:
resource_uri: URI of the resource to read
Returns:
str: Resource content
"""
try:
async with self.client as client:
result = await client.read_resource(resource_uri)
return result
except Exception as e:
print(f"❌ Error reading resource {resource_uri}: {str(e)}")
return None
async def process_query(self, query: str) -&gt; str:
"""
Process a user query, interacting with Claude and FastMCP tools and resources.
Args:
query: User query
Returns:
str: Final processed response
"""
try:
# Use FastMCP context for all operations
async with self.client as client:
# Get available tools and resources
tools_list = await client.list_tools()
resources_list = await client.list_resources()
# Prepare tools for Claude in correct format
claude_tools = []
for tool in tools_list:
claude_tool = {
"name": tool.name,
"description": tool.description or f"Tool {tool.name}",
"input_schema": tool.inputSchema or {"type": "object", "properties": {}}
}
claude_tools.append(claude_tool)
# Add a special tool for reading resources (including template resources)
resource_description = "Read a resource from the MCP server. "
if resources_list:
# Convert URIs to strings to avoid AnyUrl object issues
resource_uris = [str(r.uri) for r in resources_list]
resource_description += f"Available static resources: {', '.join(resource_uris)}. "
resource_description += "Also supports template resources like github://repo/owner/repo_name for GitHub repository information."
claude_tools.append({
"name": "read_mcp_resource",
"description": resource_description,
"input_schema": {
"type": "object",
"properties": {
"resource_uri": {
"type": "string",
"description": "URI of the resource to read. Can be static (like resource://server_info) or template-based (like github://repo/facebook/react)"
}
},
"required": ["resource_uri"]
}
})
# Create initial message for Claude
messages = [
{
"role": "user",
"content": query
}
]
# First call to Claude
response = self.anthropic.messages.create(
model="claude-3-5-sonnet-20241022",
max_tokens=6000,
messages=messages,
tools=claude_tools if claude_tools else None
)
# Process Claude's response
response_text = ""
for content_block in response.content:
if content_block.type == "text":
response_text += content_block.text
elif content_block.type == "tool_use":
# Claude wants to use a tool
tool_name = content_block.name
tool_args = content_block.input
tool_call_id = content_block.id
print(f"🔧 Claude wants to use: {tool_name}")
print(f"📝 Arguments: {tool_args}")
try:
if tool_name == "read_mcp_resource":
# Handle resource reading
resource_uri = tool_args.get("resource_uri")
if resource_uri:
tool_result = await client.read_resource(resource_uri)
print(f"📖 Resource read successfully: {resource_uri}")
# Better handling of resource result
if hasattr(tool_result, 'content'):
# If it's a resource response object, extract content
if hasattr(tool_result.content, 'text'):
result_content = tool_result.content.text
else:
result_content = str(tool_result.content)
else:
# If it's already a string or simple object
result_content = str(tool_result)
else:
tool_result = "Error: No resource URI provided"
result_content = tool_result
else:
# Execute regular tool on the FastMCP server
tool_result = await client.call_tool(tool_name, tool_args)
print(f"✅ Tool executed successfully")
result_content = str(tool_result)
# Add tool result to the conversation
messages.append({
"role": "assistant",
"content": response.content
})
# Format result for Claude
if tool_result:
messages.append({
"role": "user",
"content": [{
"type": "tool_result",
"tool_use_id": tool_call_id,
"content": f"Tool result: {result_content}"
}]
})
else:
messages.append({
"role": "user",
"content": [{
"type": "tool_result",
"tool_use_id": tool_call_id,
"content": "Tool executed without response content"
}]
})
# Second call to Claude with the tool result
final_response = self.anthropic.messages.create(
model="claude-3-5-sonnet-20241022",
max_tokens=6000,
messages=messages,
tools=claude_tools if claude_tools else None
)
# Extract text from the final response
for final_content in final_response.content:
if final_content.type == "text":
response_text += final_content.text
except Exception as e:
error_msg = f"❌ Error executing {tool_name}: {str(e)}"
print(error_msg)
response_text += f" {error_msg}"
return response_text
except Exception as e:
error_msg = f"❌ Error processing query: {str(e)}"
print(error_msg)
return error_msg
async def chat_loop(self):
"""
Main chat loop with user interaction.
"""
print(" 🤖 FastMCP client started. Write 'quit', 'q', 'exit', 'salir' to exit.")
print("💬 You can ask questions about GitHub repositories!")
print("📚 The client can use tools and resources from the FastMCP server")
print("-" * 60)
while True:
try:
# Request user input
user_input = input(" 👤 You: ").strip()
if user_input.lower() in ['quit', 'q', 'exit', 'salir']:
print("👋 Bye!")
break
if not user_input:
continue
print(" 🤔 Claude is thinking...")
# Process query
response = await self.process_query(user_input)
# Show response
print(f" 🤖 Claude: {response}")
except KeyboardInterrupt:
print(" 👋 Disconnecting...")
break
except Exception as e:
print(f" ❌ Error in chat: {str(e)}")
continue
async def cleanup(self):
"""Clean up resources and close connections."""
print("🧹 Cleaning up resources...")
# FastMCP Client cleanup is handled automatically by context manager
await self.exit_stack.aclose()
print("✅ Resources released")
async def main():
"""
Main function that initializes and runs the FastMCP client.
"""
# Verify command line arguments
if len(sys.argv) != 2:
print("❌ Usage: python client.py &lt;path_to_fastmcp_server&gt;")
print("📝 Example: python client.py ../MCP_github/github_server.py")
sys.exit(1)
server_script_path = sys.argv[1]
# Create and run client
client = FastMCPClient()
try:
# Connect to the server
await client.connect_to_server(server_script_path)
# List available tools and resources after connection
await client.list_available_tools()
await client.list_available_resources()
# Start chat loop
await client.chat_loop()
except Exception as e:
print(f"❌ Fatal error: {str(e)}")
finally:
# Ensure resources are cleaned up
await client.cleanup()
if __name__ == "__main__":
# Entry point of the script
asyncio.run(main())
Copied
>_ Output
			
Overwriting client_MCP/client.py

Como vemos le decimos "description": "URI do recurso a ser lido. Pode ser estático (como resource://server_info) ou baseado em modelo (como github://repo/facebook/react)"

claude_tools.append({
"name": "read_mcp_resource",
"description": "Ler um recurso do servidor MCP. Recursos disponíveis: " +
", ".join(resource_uris),
"input_schema": {
"type": "object",
"propriedades": {
"resource_uri": { "type": "string",
"description": "URI do recurso a ler"
}
},
"required": ["resource_uri"]
}
})

Teste do resource templatelink image 31

Executamos o cliente

	
< > Input
Python
!cd client_MCP && source .venv/bin/activate && python client.py ../gitHub_MCP_server/github_server.py
Copied
>_ Output
			
🔗 Connecting to FastMCP server: ../gitHub_MCP_server/github_server.py
✅ Client created successfully
🛠️ Available tools (2):
==================================================
📋 sub_mcp_hello_world
Description: Returns a simple greeting.
Parameters:
📋 list_repository_issues
Description: Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
ctx: The context of the request
user_id: The user ID (automatically injected by the server)
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
Parameters: owner, repo_name
📚 Available resources (1):
==================================================
📄 resource://server_info
...
- react
- ui
The repository can be accessed via:
- HTTPS: https://github.com/facebook/react
- SSH: git@github.com:facebook/react.git
This is one of the most popular repositories on GitHub, as evidenced by its high number of stars and forks, and it remains actively maintained with regular updates.
👤 You: q
👋 Bye!
🧹 Cleaning up resources...
✅ Resources released

Pedimos a informação de um repositório, use o resource template repository_info e ele nos fornece as informações do repositório.

👤 Você: Pode ler o recurso github://repo/facebook/react para obter informações detalhadas sobre o repositório?

🤔 Claude está pensando...
🔧 Claude quer usar: read_mcp_resource
📝 Argumentos: {'resource_uri': 'github://repo/facebook/react'}
📖 Recurso lido com sucesso: github://repo/facebook/react

🤖 Claude: Vou ajudar você a ler as informações do repositório GitHub do React da Facebook usando a função `read_mcp_resource`.Com base nas informações do repositório recuperadas, aqui estão os principais detalhes sobre o repositório React da Facebook:

1. Descrição: A biblioteca para interfaces de usuário web e nativas
2. Proprietário: Facebook (Organização)
3. Linguagem: JavaScript
4. Estatísticas do Repositório:
- Estrelas: 236,803
- Forks: 48,815
- Questões em aberto: 999
- Observadores: 236,803

5. Datas Importantes:
- Criado: 24 de maio de 2013
- Última atualização: 28 de junho de 2025
- Último push: 27 de junho de 2025

6. Recursos do repositório:
- Repositório público (não privado)
- Não é um fork- Não arquivado
- Tem issues habilitados
- Projetos desativados
- Wiki desativada

7. Licença: MIT License

8. Tópicos/Tags:
- declarativo
- frontend
- javascript
- biblioteca
- react
- ui

O repositório pode ser acessado via:
- HTTPS: https://github.com/facebook/react
- SSH: git@github.com:facebook/react.git

Este é um dos repositórios mais populares no GitHub, como evidenciado pelo seu alto número de estrelas e forks, e continua sendo mantido ativamente com atualizações regulares.

Criar um promptlink image 32

Outra das ferramentas que o MCP nos oferece é pedir ao LLM que nos crie um prompt para usá-lo em uma solicitação.

Servidor MCPlink image 33

Criamos um prompt em nosso servidor; para isso, usamos o decorador @mcp.prompt e passamos o nome do prompt, a descrição e a etiqueta public, porque tínhamos definido nosso servidor para que incluísse apenas as tools, os resources e os prompts com a etiqueta public.

@mcp.prompt(
name="generate_issues_prompt",
description="Gera um prompt estruturado para perguntar sobre issues de repositórios do GitHub. Use isto quando os usuários quiserem formular perguntas sobre issues de repositórios, ou precisarem de ajuda para criar prompts para análise de issues.",tags={"público"}
)
	
< > Input
Python
%%writefile gitHub_MCP_server/github_server.py
import httpx
from fastmcp import FastMCP, Context
from github import GITHUB_TOKEN, create_github_headers
import datetime
USER_ID = 1234567890
# Create FastMCP server
mcp = FastMCP(
name="GitHubMCP",
instructions="This server provides tools, resources and prompts to interact with the GitHub API.",
include_tags={"public"},
exclude_tags={"first_issue"}
)
sub_mcp = FastMCP(
name="SubMCP",
)
@mcp.tool(
tags={"public", "production"},
exclude_args=["user_id"], # user_id has to be injected by server, not provided by LLM
)
async def list_repository_issues(owner: str, repo_name: str, ctx: Context, user_id: int = USER_ID) -&gt; list[dict]:
"""
Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
ctx: The context of the request
user_id: The user ID (automatically injected by the server)
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
"""
# Limit to first 10 issues to avoid very long responses
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=open&amp;per_page=10"
ctx.info(f"Fetching issues from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
ctx.info("No open issues found for this repository.")
return [{"message": "No open issues found for this repository."}]
issues_summary = []
for issue in issues_data:
# Create a more concise summary
summary = f"#{issue.get('number', 'N/A')}: {issue.get('title', 'No title')}"
if issue.get('comments', 0) &gt; 0:
summary += f" ({issue.get('comments')} comments)"
issues_summary.append({
"number": issue.get("number"),
"title": issue.get("title"),
"user": issue.get("user", {}).get("login"),
"url": issue.get("html_url"),
"comments": issue.get("comments"),
"summary": summary
})
ctx.info(f"Found {len(issues_summary)} open issues.")
# Add context information
result = {
"total_found": len(issues_summary),
"repository": f"{owner}/{repo_name}",
"note": "Showing first 10 open issues" if len(issues_summary) == 10 else f"Showing all {len(issues_summary)} open issues",
"issues": issues_summary,
"requested_by_user_id": user_id
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
ctx.error(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
ctx.error(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.tool(tags={"private", "development"})
async def list_more_repository_issues(owner: str, repo_name: str) -&gt; list[dict]:
"""
Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
"""
# Limit to first 100 issues to avoid very long responses
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=open&amp;per_page=100"
print(f"Fetching issues from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
print("No open issues found for this repository.")
return [{"message": "No open issues found for this repository."}]
issues_summary = []
for issue in issues_data:
# Create a more concise summary
summary = f"#{issue.get('number', 'N/A')}: {issue.get('title', 'No title')}"
if issue.get('comments', 0) &gt; 0:
summary += f" ({issue.get('comments')} comments)"
issues_summary.append({
"number": issue.get("number"),
"title": issue.get("title"),
"user": issue.get("user", {}).get("login"),
"url": issue.get("html_url"),
"comments": issue.get("comments"),
"summary": summary
})
print(f"Found {len(issues_summary)} open issues.")
# Add context information
result = {
"total_found": len(issues_summary),
"repository": f"{owner}/{repo_name}",
"note": "Showing first 10 open issues" if len(issues_summary) == 10 else f"Showing all {len(issues_summary)} open issues",
"issues": issues_summary
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
print(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
print(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.tool(tags={"public", "first_issue"})
async def first_repository_issue(owner: str, repo_name: str) -&gt; list[dict]:
"""
Gets the first issue ever created in a GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
Returns:
list[dict]: A list containing information about the first issue created
"""
# Get the first issue by sorting by creation date in ascending order
api_url = f"https://api.github.com/repos/{owner}/{repo_name}/issues?state=all&amp;sort=created&amp;direction=asc&amp;per_page=1"
print(f"Fetching first issue from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
issues_data = response.json()
if not issues_data:
print("No issues found for this repository.")
return [{"message": "No issues found for this repository."}]
first_issue = issues_data[0]
# Create a detailed summary of the first issue
summary = f"#{first_issue.get('number', 'N/A')}: {first_issue.get('title', 'No title')}"
if first_issue.get('comments', 0) &gt; 0:
summary += f" ({first_issue.get('comments')} comments)"
issue_info = {
"number": first_issue.get("number"),
"title": first_issue.get("title"),
"user": first_issue.get("user", {}).get("login"),
"url": first_issue.get("html_url"),
"state": first_issue.get("state"),
"comments": first_issue.get("comments"),
"created_at": first_issue.get("created_at"),
"updated_at": first_issue.get("updated_at"),
"body": first_issue.get("body", "")[:500] + "..." if len(first_issue.get("body", "")) &gt; 500 else first_issue.get("body", ""),
"summary": summary
}
print(f"Found first issue: #{first_issue.get('number')} created on {first_issue.get('created_at')}")
# Add context information
result = {
"repository": f"{owner}/{repo_name}",
"note": "This is the very first issue created in this repository",
"first_issue": issue_info
}
return [result]
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
print(f"GitHub API error: {e.response.status_code}. {error_message}")
return [{
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message
}]
except Exception as e:
print(f"An unexpected error occurred: {str(e)}")
return [{"error": f"An unexpected error occurred: {str(e)}"}]
@mcp.resource("resource://server_info", tags={"public"})
def server_info(ctx: Context) -&gt; str:
"""
Returns information about the server.
"""
return {
"info": "This is the MCP GitHub server development for MaximoFN blog post",
"requested_id": ctx.request_id
}
# Use: ¿Puedes leer el resource github://repo/facebook/react para obtener información detallada del repositorio?
@mcp.resource("github://repo/{owner}/{repo_name}", tags={"public"})
async def repository_info(owner: str, repo_name: str, ctx: Context) -&gt; dict:
"""
Returns detailed information about a GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
ctx: The context of the request
Returns:
dict: Repository information including name, description, stats, etc.
"""
api_url = f"https://api.github.com/repos/{owner}/{repo_name}"
ctx.info(f"Fetching repository information from {api_url}...")
async with httpx.AsyncClient() as client:
try:
response = await client.get(api_url, headers=create_github_headers())
response.raise_for_status()
repo_data = response.json()
# Extract relevant repository information
repo_info = {
"name": repo_data.get("name"),
"full_name": repo_data.get("full_name"),
"description": repo_data.get("description"),
"owner": {
"login": repo_data.get("owner", {}).get("login"),
"type": repo_data.get("owner", {}).get("type")
},
"html_url": repo_data.get("html_url"),
"clone_url": repo_data.get("clone_url"),
"ssh_url": repo_data.get("ssh_url"),
"language": repo_data.get("language"),
"size": repo_data.get("size"), # Size in KB
"stargazers_count": repo_data.get("stargazers_count"),
"watchers_count": repo_data.get("watchers_count"),
"forks_count": repo_data.get("forks_count"),
"open_issues_count": repo_data.get("open_issues_count"),
"default_branch": repo_data.get("default_branch"),
"created_at": repo_data.get("created_at"),
"updated_at": repo_data.get("updated_at"),
"pushed_at": repo_data.get("pushed_at"),
"is_private": repo_data.get("private"),
"is_fork": repo_data.get("fork"),
"is_archived": repo_data.get("archived"),
"has_issues": repo_data.get("has_issues"),
"has_projects": repo_data.get("has_projects"),
"has_wiki": repo_data.get("has_wiki"),
"license": repo_data.get("license", {}).get("name") if repo_data.get("license") else None,
"topics": repo_data.get("topics", [])
}
ctx.info(f"Successfully retrieved information for repository {owner}/{repo_name}")
return repo_info
except httpx.HTTPStatusError as e:
error_message = e.response.json().get("message", "No additional message from API.")
if e.response.status_code == 404:
error_message = f"Repository {owner}/{repo_name} not found or is private."
elif e.response.status_code == 403 and GITHUB_TOKEN:
error_message += " (Rate limit with token or token lacks permissions?)"
elif e.response.status_code == 403 and not GITHUB_TOKEN:
error_message += " (Rate limit without token. Consider creating a .env file with GITHUB_TOKEN.)"
ctx.error(f"GitHub API error: {e.response.status_code}. {error_message}")
return {
"error": f"GitHub API error: {e.response.status_code}",
"message": error_message,
"repository": f"{owner}/{repo_name}"
}
except Exception as e:
ctx.error(f"An unexpected error occurred: {str(e)}")
return {
"error": f"An unexpected error occurred: {str(e)}",
"repository": f"{owner}/{repo_name}"
}
@mcp.prompt(
name="generate_issues_prompt",
description="Generates a structured prompt for asking about GitHub repository issues. Use this when users want to formulate questions about repository issues, or need help creating prompts for issue analysis.",
tags={"public"}
)
def generate_issues_prompt(owner: str, repo_name: str) -&gt; str:
"""
Generates a structured prompt for asking about GitHub repository issues.
This prompt template helps users formulate clear questions about repository issues
and can be used as a starting point for issue analysis or research.
Args:
owner: Repository owner (e.g., 'huggingface', 'microsoft')
repo_name: Repository name (e.g., 'transformers', 'vscode')
Returns:
A formatted prompt asking about repository issues
"""
return f"""Please provide information about the open issues in the repository {owner}/{repo_name}.
I'm interested in:
- Current open issues and their status
- Recent issue trends and patterns
- Common issue categories or topics
- Any critical or high-priority issues
Repository: {owner}/{repo_name}"""
@sub_mcp.tool(tags={"public"})
def hello_world() -&gt; str:
"""
Returns a simple greeting.
"""
return "Hello, world!"
mcp.mount("sub_mcp", sub_mcp)
if __name__ == "__main__":
print("DEBUG: Starting FastMCP GitHub server...")
print(f"DEBUG: Server name: {mcp.name}")
# Initialize and run the server
mcp.run(
transport="stdio"
)
Copied
>_ Output
			
Overwriting gitHub_MCP_server/github_server.py

Cliente MCPlink image 34

Modificamos nosso cliente para poder usar o prompt que criamos em nosso servidor.

	
< > Input
Python
%%writefile client_MCP/client.py
import sys
import asyncio
from contextlib import AsyncExitStack
from anthropic import Anthropic
from dotenv import load_dotenv
from fastmcp import Client
# Load environment variables from .env file
load_dotenv()
class FastMCPClient:
"""
FastMCP client that integrates with Claude to process user queries
and use tools and resources exposed by a FastMCP server.
"""
def __init__(self):
"""Initialize the FastMCP client with Anthropic and resource management."""
self.exit_stack = AsyncExitStack()
self.anthropic = Anthropic()
self.client = None
async def connect_to_server(self, server_script_path: str):
"""
Connect to the specified FastMCP server.
Args:
server_script_path: Path to the server script (Python)
"""
print(f"🔗 Connecting to FastMCP server: {server_script_path}")
# Determine the server type based on the extension
if not server_script_path.endswith('.py'):
raise ValueError(f"Unsupported server type. Use .py files. Received: {server_script_path}")
# Create FastMCP client
self.client = Client(server_script_path)
# Note: FastMCP Client automatically infers transport from .py files
print("✅ Client created successfully")
async def list_available_tools(self):
"""List available tools in the FastMCP server."""
try:
# Get list of tools from the server using FastMCP context
async with self.client as client:
tools = await client.list_tools()
if tools:
print(f" 🛠️ Available tools ({len(tools)}):")
print("=" * 50)
for tool in tools:
print(f"📋 {tool.name}")
if tool.description:
print(f" Description: {tool.description}")
# Show parameters if available
if hasattr(tool, 'inputSchema') and tool.inputSchema:
if 'properties' in tool.inputSchema:
params = list(tool.inputSchema['properties'].keys())
print(f" Parameters: {', '.join(params)}")
print()
else:
print("⚠️ No tools found in the server")
except Exception as e:
print(f"❌ Error listing tools: {str(e)}")
async def list_available_resources(self):
"""List available resources in the FastMCP server."""
try:
# Get list of resources from the server using FastMCP context
async with self.client as client:
resources = await client.list_resources()
if resources:
print(f" 📚 Available resources ({len(resources)}):")
print("=" * 50)
for resource in resources:
print(f"📄 {resource.uri}")
if resource.name:
print(f" Name: {resource.name}")
if resource.description:
print(f" Description: {resource.description}")
if resource.mimeType:
print(f" MIME Type: {resource.mimeType}")
print()
else:
print("⚠️ No resources found in the server")
except Exception as e:
print(f"❌ Error listing resources: {str(e)}")
async def list_available_prompts(self):
"""List available prompts in the FastMCP server."""
try:
# Get list of prompts from the server using FastMCP context
async with self.client as client:
prompts = await client.list_prompts()
if prompts:
print(f" 💭 Available prompts ({len(prompts)}):")
print("=" * 50)
for prompt in prompts:
print(f"🎯 {prompt.name}")
if prompt.description:
print(f" Description: {prompt.description}")
# Show parameters if available
if hasattr(prompt, 'arguments') and prompt.arguments:
params = []
for arg in prompt.arguments:
param_info = f"{arg.name}: {arg.description or 'No description'}"
if arg.required:
param_info += " (required)"
params.append(param_info)
print(f" Parameters: {', '.join(params)}")
print()
else:
print("⚠️ No prompts found in the server")
except Exception as e:
print(f"❌ Error listing prompts: {str(e)}")
async def read_resource(self, resource_uri: str):
"""
Read a specific resource from the server.
Args:
resource_uri: URI of the resource to read
Returns:
str: Resource content
"""
try:
async with self.client as client:
result = await client.read_resource(resource_uri)
return result
except Exception as e:
print(f"❌ Error reading resource {resource_uri}: {str(e)}")
return None
async def get_prompt(self, prompt_name: str, prompt_args: dict = None):
"""
Get/call a specific prompt from the server.
Args:
prompt_name: Name of the prompt to call
prompt_args: Arguments for the prompt (if any)
Returns:
str: Generated prompt content
"""
try:
async with self.client as client:
if prompt_args:
result = await client.get_prompt(prompt_name, prompt_args)
else:
result = await client.get_prompt(prompt_name)
# Extract the prompt text from the response
if hasattr(result, 'messages') and result.messages:
# FastMCP returns prompts as message objects
return ' '.join([msg.content.text for msg in result.messages if hasattr(msg.content, 'text')])
elif hasattr(result, 'content'):
return str(result.content)
else:
return str(result)
except Exception as e:
print(f"❌ Error getting prompt {prompt_name}: {str(e)}")
return None
async def process_query(self, query: str) -&gt; str:
"""
Process a user query, interacting with Claude and FastMCP tools and resources.
Args:
query: User query
Returns:
str: Final processed response
"""
try:
# Use FastMCP context for all operations
async with self.client as client:
# Get available tools, resources, and prompts
tools_list = await client.list_tools()
resources_list = await client.list_resources()
prompts_list = await client.list_prompts()
# Prepare tools for Claude in correct format
claude_tools = []
for tool in tools_list:
claude_tool = {
"name": tool.name,
"description": tool.description or f"Tool {tool.name}",
"input_schema": tool.inputSchema or {"type": "object", "properties": {}}
}
claude_tools.append(claude_tool)
# Add a special tool for reading resources (including template resources)
resource_description = "Read a resource from the MCP server. "
if resources_list:
# Convert URIs to strings to avoid AnyUrl object issues
resource_uris = [str(r.uri) for r in resources_list]
resource_description += f"Available static resources: {', '.join(resource_uris)}. "
resource_description += "Also supports template resources like github://repo/owner/repo_name for GitHub repository information."
claude_tools.append({
"name": "read_mcp_resource",
"description": resource_description,
"input_schema": {
"type": "object",
"properties": {
"resource_uri": {
"type": "string",
"description": "URI of the resource to read. Can be static (like resource://server_info) or template-based (like github://repo/facebook/react)"
}
},
"required": ["resource_uri"]
}
})
# Add a special tool for using prompts
prompt_description = "Generate specialized prompts from the MCP server. Use this when users want to: "
prompt_description += "- Create well-structured questions about repositories "
prompt_description += "- Get help formulating prompts for specific tasks "
prompt_description += "- Generate template questions for analysis "
if prompts_list:
prompt_names = [p.name for p in prompts_list]
prompt_description += f" Available prompts: {', '.join(prompt_names)} "
prompt_description += "- generate_issues_prompt: Creates structured questions about GitHub repository issues"
prompt_description += " IMPORTANT: Use prompts when users explicitly ask for help creating questions or prompts, or when they want to formulate better questions about repositories."
claude_tools.append({
"name": "use_mcp_prompt",
"description": prompt_description,
"input_schema": {
"type": "object",
"properties": {
"prompt_name": {
"type": "string",
"description": "Name of the prompt to use. Available: 'generate_issues_prompt'"
},
"prompt_args": {
"type": "object",
"description": "Arguments for the prompt. For generate_issues_prompt: {'owner': 'repo-owner', 'repo_name': 'repo-name'}",
"properties": {
"owner": {
"type": "string",
"description": "Repository owner (e.g., 'huggingface', 'microsoft')"
},
"repo_name": {
"type": "string",
"description": "Repository name (e.g., 'transformers', 'vscode')"
}
}
}
},
"required": ["prompt_name"]
}
})
# Create initial message for Claude
messages = [
{
"role": "user",
"content": query
}
]
# First call to Claude
response = self.anthropic.messages.create(
model="claude-3-5-sonnet-20241022",
max_tokens=6000,
messages=messages,
tools=claude_tools if claude_tools else None
)
# Process Claude's response
response_text = ""
for content_block in response.content:
if content_block.type == "text":
response_text += content_block.text
elif content_block.type == "tool_use":
# Claude wants to use a tool
tool_name = content_block.name
tool_args = content_block.input
tool_call_id = content_block.id
print(f"🔧 Claude wants to use: {tool_name}")
print(f"📝 Arguments: {tool_args}")
try:
if tool_name == "read_mcp_resource":
# Handle resource reading
resource_uri = tool_args.get("resource_uri")
if resource_uri:
tool_result = await client.read_resource(resource_uri)
print(f"📖 Resource read successfully: {resource_uri}")
# Better handling of resource result
if hasattr(tool_result, 'content'):
# If it's a resource response object, extract content
if hasattr(tool_result.content, 'text'):
result_content = tool_result.content.text
else:
result_content = str(tool_result.content)
else:
# If it's already a string or simple object
result_content = str(tool_result)
else:
tool_result = "Error: No resource URI provided"
result_content = tool_result
elif tool_name == "use_mcp_prompt":
# Handle prompt usage
prompt_name = tool_args.get("prompt_name")
prompt_args = tool_args.get("prompt_args", {})
if prompt_name:
tool_result = await self.get_prompt(prompt_name, prompt_args)
print(f"💭 Prompt '{prompt_name}' generated successfully")
result_content = str(tool_result) if tool_result else "Error generating prompt"
else:
tool_result = "Error: No prompt name provided"
result_content = tool_result
else:
# Execute regular tool on the FastMCP server
tool_result = await client.call_tool(tool_name, tool_args)
print(f"✅ Tool executed successfully")
result_content = str(tool_result)
# Add tool result to the conversation
messages.append({
"role": "assistant",
"content": response.content
})
# Format result for Claude
if tool_result:
messages.append({
"role": "user",
"content": [{
"type": "tool_result",
"tool_use_id": tool_call_id,
"content": f"Tool result: {result_content}"
}]
})
else:
messages.append({
"role": "user",
"content": [{
"type": "tool_result",
"tool_use_id": tool_call_id,
"content": "Tool executed without response content"
}]
})
# Second call to Claude with the tool result
final_response = self.anthropic.messages.create(
model="claude-3-5-sonnet-20241022",
max_tokens=6000,
messages=messages,
tools=claude_tools if claude_tools else None
)
# Extract text from the final response
for final_content in final_response.content:
if final_content.type == "text":
response_text += final_content.text
except Exception as e:
error_msg = f"❌ Error executing {tool_name}: {str(e)}"
print(error_msg)
response_text += f" {error_msg}"
return response_text
except Exception as e:
error_msg = f"❌ Error processing query: {str(e)}"
print(error_msg)
return error_msg
async def chat_loop(self):
"""
Main chat loop with user interaction.
"""
print(" 🤖 FastMCP client started. Write 'quit', 'q', 'exit', 'salir' to exit.")
print("💬 You can ask questions about GitHub repositories!")
print("📚 The client can use tools, resources, and prompts from the FastMCP server")
print()
print("💭 PROMPT Examples:")
print(" • 'Generate a prompt for asking about issues in facebook/react'")
print(" • 'Help me create a good question about microsoft/vscode issues'")
print(" • 'I need a structured prompt for analyzing tensorflow/tensorflow'")
print()
print("🔧 DIRECT Examples:")
print(" • 'Show me the issues in huggingface/transformers'")
print(" • 'Get repository info for github://repo/google/chrome'")
print("-" * 60)
while True:
try:
# Request user input
user_input = input(" 👤 You: ").strip()
if user_input.lower() in ['quit', 'q', 'exit', 'salir']:
print("👋 Bye!")
break
if not user_input:
continue
print(" 🤔 Claude is thinking...")
# Process query
response = await self.process_query(user_input)
# Show response
print(f" 🤖 Claude: {response}")
except KeyboardInterrupt:
print(" 👋 Disconnecting...")
break
except Exception as e:
print(f" ❌ Error in chat: {str(e)}")
continue
async def cleanup(self):
"""Clean up resources and close connections."""
print("🧹 Cleaning up resources...")
# FastMCP Client cleanup is handled automatically by context manager
await self.exit_stack.aclose()
print("✅ Resources released")
async def main():
"""
Main function that initializes and runs the FastMCP client.
"""
# Verify command line arguments
if len(sys.argv) != 2:
print("❌ Usage: python client.py &lt;path_to_fastmcp_server&gt;")
print("📝 Example: python client.py ../MCP_github/github_server.py")
sys.exit(1)
server_script_path = sys.argv[1]
# Create and run client
client = FastMCPClient()
try:
# Connect to the server
await client.connect_to_server(server_script_path)
# List available tools, resources, and prompts after connection
await client.list_available_tools()
await client.list_available_resources()
await client.list_available_prompts()
# Start chat loop
await client.chat_loop()
except Exception as e:
print(f"❌ Fatal error: {str(e)}")
finally:
# Ensure resources are cleaned up
await client.cleanup()
if __name__ == "__main__":
# Entry point of the script
asyncio.run(main())
Copied
>_ Output
			
Overwriting client_MCP/client.py

Criamos as funções list_available_prompts e get_prompt para listar os prompts disponíveis e obter um prompt específico.

Teste do promptlink image 35

Executamos o cliente

	
< > Input
Python
!cd client_MCP && source .venv/bin/activate && python client.py ../gitHub_MCP_server/github_server.py
Copied
>_ Output
			
🔗 Connecting to FastMCP server: ../gitHub_MCP_server/github_server.py
✅ Client created successfully
🛠️ Available tools (2):
==================================================
📋 sub_mcp_hello_world
Description: Returns a simple greeting.
Parameters:
📋 list_repository_issues
Description: Lists open issues for a given GitHub repository.
Args:
owner: The owner of the repository (e.g., 'modelcontextprotocol')
repo_name: The name of the repository (e.g., 'python-sdk')
ctx: The context of the request
user_id: The user ID (automatically injected by the server)
Returns:
list[dict]: A list of dictionaries, each containing information about an issue
Parameters: owner, repo_name
📚 Available resources (1):
==================================================
📄 resource://server_info
...
💭 Prompt 'generate_issues_prompt' generated successfully
🤖 Claude: I'll help you generate a structured prompt for viewing issues from the Hugging Face Transformers repository using the `use_mcp_prompt` function with the `generate_issues_prompt` prompt type. I have all the required information from your request:
- owner: `huggingface'
- repo_name: 'transformers'
I've generated a structured prompt that you can use to analyze issues in the Hugging Face Transformers repository. This prompt is designed to help you get comprehensive information about the repository's issues, including their current status, trends, categories, and priorities.
Would you like me to actually fetch the current issues from the repository using this prompt? If so, I can use the `list_repository_issues` function to get that information for you.
👤 You: q
👋 Bye!
🧹 Cleaning up resources...
✅ Resources released

Vemos que nos dá uma lista dos prompts disponíveis.

💭 Prompts disponíveis (1):
==================================================
🎯 generate_issues_prompt
Descrição: Gera um prompt estruturado para perguntar sobre issues de um repositório no GitHub. Use isto quando os utilizadores quiserem formular perguntas sobre issues do repositório ou precisarem de ajuda para criar prompts para análise de issues.
Parâmetros: owner: Sem descrição (obrigatório), repo_name: Sem descrição (obrigatório)


🤖 FastMCP cliente iniciado. Escreva 'quit', 'q', 'exit', 'salir' para sair.
💬 Você pode fazer perguntas sobre repositórios do GitHub!
📚 O cliente pode usar ferramentas, recursos e prompts do servidor FastMCP

💭 Exemplos de PROMPT:
• 'Gere um prompt para perguntar sobre problemas no facebook/react'
• 'Ajude-me a criar uma boa pergunta sobre problemas do microsoft/vscode'
• 'Preciso de um prompt estruturado para analisar tensorflow/tensorflow'

🔧 Exemplos DIRETOS:
• 'Mostre-me os problemas em huggingface/transformers'
• 'Obter informações do repositório para github://repo/google/chrome'

E que, quando pedimos um prompt, ele nos dá o prompt gerado.

👤 Você: Você pode criar um prompt para visualizar issues do repositório Hugging Face Transformers?

🤔 Claude está pensando...
🔧 Claude quer usar: use_mcp_prompt
📝 Argumentos: {'prompt_name': 'generate_issues_prompt', 'prompt_args': {'owner': 'huggingface', 'repo_name': 'transformers'}}
💭 Prompt `generate_issues_prompt` gerado com sucesso

🤖 Claude: Vou ajudar você a gerar um prompt estruturado para visualizar issues do repositório Hugging Face Transformers usando a função `use_mcp_prompt` com o tipo de prompt `generate_issues_prompt`. Tenho todas as informações necessárias do seu pedido:
- proprietário: "huggingface"
- repo_name: "transformers"Eu gerei um prompt estruturado que você pode usar para analisar issues no repositório Hugging Face Transformers. Este prompt foi projetado para ajudar você a obter informações abrangentes sobre as issues do repositório, incluindo seu status atual, tendências, categorias e prioridades.

Você gostaria que eu realmente buscasse os issues atuais do repositório usando este prompt? Se sim, posso usar a função `list_repository_issues` para obter essas informações para você.

---

➡️ **Continua na Parte 4: HTTP, autenticação e cliente MCP**.

Continuar lendo

Últimos posts -->

Você viu esses projetos?

Gymnasia

Gymnasia Gymnasia
React Native
Expo
TypeScript
FastAPI
Next.js
OpenAI
Anthropic

Aplicativo móvel de treino pessoal com assistente de IA, biblioteca de exercícios, acompanhamento de rotinas, dieta e medidas corporais

Horeca chatbot

Horeca chatbot Horeca chatbot
Python
LangChain
PostgreSQL
PGVector
React
Kubernetes
Docker
GitHub Actions

Chatbot conversacional para cozinheiros de hotéis e restaurantes. Um cozinheiro, gerente de cozinha ou serviço de quarto de um hotel ou restaurante pode falar com o chatbot para obter informações sobre receitas e menus. Mas também implementa agentes, com os quais pode editar ou criar novas receitas ou menus

Naviground

Naviground Naviground
Ver todos os projetos -->
>_ Disponível para projetos

Tem um projeto com IA?

Vamos conversar.

maximofn@gmail.com

Especialista em Machine Learning e Inteligência Artificial. Desenvolvo soluções com IA generativa, agentes inteligentes e modelos personalizados.

Quer assistir alguma palestra?

Últimas palestras -->

Quer melhorar com essas dicas?

Últimos tips -->

Use isso localmente

Os espaços do Hugging Face nos permitem executar modelos com demos muito simples, mas e se a demo quebrar? Ou se o usuário a deletar? Por isso, criei contêineres docker com alguns espaços interessantes, para poder usá-los localmente, aconteça o que acontecer. Na verdade, se você clicar em qualquer botão de visualização de projeto, ele pode levá-lo a um espaço que não funciona.

Flow edit

Flow edit Flow edit

Edite imagens com este modelo de Flow. Baseado em SD3 ou FLUX, você pode editar qualquer imagem e gerar novas

FLUX.1-RealismLora

FLUX.1-RealismLora FLUX.1-RealismLora
Ver todos os contêineres -->
>_ Disponível para projetos

Tem um projeto com IA?

Vamos conversar.

maximofn@gmail.com

Especialista em Machine Learning e Inteligência Artificial. Desenvolvo soluções com IA generativa, agentes inteligentes e modelos personalizados.

Você quer treinar seu modelo com esses datasets?

short-jokes-dataset

HuggingFace

Dataset com piadas em inglês

Uso: Fine-tuning de modelos de geração de texto humorístico

231K linhas 2 colunas 45 MB
Ver no HuggingFace →

opus100

HuggingFace

Dataset com traduções de inglês para espanhol

Uso: Treinamento de modelos de tradução inglês-espanhol

1M linhas 2 colunas 210 MB
Ver no HuggingFace →

netflix_titles

HuggingFace

Dataset com filmes e séries da Netflix

Uso: Análise de catálogo Netflix e sistemas de recomendação

8.8K linhas 12 colunas 3.5 MB
Ver no HuggingFace →
Ver mais datasets -->