agent · Perplexity MCP Server MCP: Risk & Policy

WHAT IT DOES

What agent does on Perplexity MCP Server

AI agents invoke agent to trigger actions in Perplexity MCP Server. 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 agent needs a policy

The 'agent' tool is an agentic system that can execute code (via sandbox/code-execution), fetch remote content, and chain multiple external operations. While it doesn't directly delete data (Destructive) or move money (Financial), it can trigger external effects through code execution and tool composition. This places it in Execute category.

From the tool's definition Tool description explicitly mentions 'agentic loop that can call built-in tools' including 'sandbox/code-execution', 'web_search', 'fetch_url', and ability to 'run third-party models'.

Attacks that exploit this kind of access

Recommended rule Ready

Verdict

agent Rate-limited

agent 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": {
    "agent": {
      "limits": [
        {
          "counter": "agent_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 Perplexity MCP Server, so you can run agents against it safely.

FAQ

Questions about agent

What does the agent tool do? +

Perplexity Agent API: an agentic loop that can call built-in tools (web_search, fetch_url, people_search, finance_search, sandbox/code-execution) and run third-party models. Use for multi-step tasks that need tool use or a specific external model. Distinct from. It is categorised as a Execute tool in the Perplexity MCP Server MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.

How do I enforce a policy on agent? +

Register the Perplexity MCP Server MCP server in PolicyLayer and add a rule for agent: 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 Perplexity MCP Server. Nothing to install.

What risk level is agent? +

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

Can I rate-limit agent? +

Yes. Add a rate_limit block to the agent 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 agent completely? +

Set action: deny in the PolicyLayer policy for agent. 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 agent? +

agent is provided by the Perplexity MCP Server MCP server (rossh121/perplexity-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.