Is there a way to limit tools available to an agent from an MCP server?

[desmondcarot]

I'm currently working with @fast.Agent and using setting it up like this:

@fast.Agent("agent1", servers=["filesystem"])

This setup works fine, but it gives the agent access to all the tools provided by the filesystem server. For security and functionality purposes, I want to restrict which tools from a server the agent can access.

Is there a way to do something like this?

@fast.Agent(
    "agent1",
    servers=[
        {
            "filesystem": {
                "tools": ["list", "of", "tools"]
            }
        }
    ]
)

Basically, I want to allow only a subset of tools from the filesystem server.

Does the @fast.Agent API (or MCP server config in general) support something like this?

Thanks in advance!

[evalstate]

This is a great suggestion.

It's ultimately down to the Host/Client what gets presented to the LLM/Agent. Thinking aloud, this feels like something that would be better configured in the fastagent.config.yaml file than in the python...

[desmondcarot]

Thanks so much for the input, and honestly, appreciate for all the work you’ve put into the project. It’s been awesome to work with so far. Totally agree that handling tool restrictions in the config file would be a cleaner approach.

[iqdoctor]

https://fast-agent.ai/agents/defining/#basic-agent

  #tools={"filesystem": ["tool_1", "tool_2"]  # Filter the tools available to the agent. Defaults to all

@fast.agent(
  name="agent",                          # name of the agent
  instruction="You are a helpful Agent", # base instruction for the agent
  servers=["filesystem"],                # list of MCP Servers for the agent
  #tools={"filesystem": ["tool_1", "tool_2"]  # Filter the tools available to the agent. Defaults to all
  #resources={"filesystem: ["resource_1", "resource_2"]} # Filter the resources available to the agent. Defaults to all
  #prompts={"filesystem": ["prompt_1", "prompt_2"]}  # Filter the prompts available to the agent. Defaults to all.
  model="o3-mini.high",                  # specify a model for the agent
  use_history=True,                      # agent maintains chat history
  request_params=RequestParams(temperature= 0.7), # additional parameters for the LLM (or RequestParams())
  human_input=True,                      # agent can request human input
  elicitation_handler=ElicitationFnT,    # custom elicitation handler (from mcp.client.session)
  api_key="programmatic-api-key",        # specify the API KEY programmatically, it will override which provided in config file or env var
)