send_input · Persistent Shell MCP: Risk & Policy

WHAT IT DOES

What send_input does on Persistent Shell MCP

AI agents invoke send_input to trigger actions in Persistent Shell MCP. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.

WHY IT NEEDS A POLICY

Why send_input needs a policy

This tool sends input to an already-running process within a persistent tmux session. Depending on what the target process is (shell, interpreter, application), arbitrary input could trigger destructive or dangerous operations. The effect is Execute since it drives external process behavior based on arguments, with high severity due to the persistent shell context where misuse could cause significant harm.

From the tool's definition Send input to a running process in a specified window

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

send_input Rate-limited

send_input stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.

View as policy code

policy.json

{
  "version": "1",
  "default": "deny",
  "tools": {
    "send_input": {
      "limits": [
        {
          "counter": "send_input_rate",
          "window": "minute",
          "max": 10,
          "scope": "grant"
        }
      ]
    }
  }
}

RATE-LIMIT THIS TOOL →

Instant setup, no code required.

PolicyLayer is an MCP gateway: every tool call is checked against your rule before it reaches Persistent Shell MCP, so you can run agents against it safely.

FAQ

Questions about send_input

What does the send_input tool do? +

Send input to a running process in a specified window. It is categorised as a Execute tool in the Persistent Shell MCP MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on send_input? +

Register the Persistent Shell MCP server in PolicyLayer and add a rule for send_input: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Persistent Shell MCP. Nothing to install.

What risk level is send_input? +

send_input is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.

Can I rate-limit send_input? +

Yes. Add a rate_limit block to the send_input rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.

How do I block send_input completely? +

Set action: deny in the PolicyLayer policy for send_input. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.

What MCP server provides send_input? +

send_input is provided by the Persistent Shell MCP server (tntisdial/persistent-shell-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.