Kilo Code QA testing, in one tool call

Kilo Code just shipped a change. Can it check that the page still works — the analytics fire, the cookies are Secure, the layout doesn't overflow on a phone — without you opening a browser?

Add the Prufa MCP server to Kilo Code and it can. The agent calls prufa_run_audit, Prufa loads the URL in a real browser, captures every network beacon, console error, cookie and response code, runs a fixed suite of deterministic checks against that captured evidence, and returns the verdict as JSON. Kilo Code reads it, fixes what broke, and runs it again. The first call needs no key and no card.

Why give Kilo Code a QA tool over MCP?

Kilo Code writes code fast and checks it badly. It can read a stack trace, but it can't see that the signup button does nothing, that the analytics tag loaded but never fired, or that the hero overflows 103px past the viewport on a phone. The usual answer — "open a browser and look" — breaks the agent loop: it needs a human, or it needs you to hand-author a Playwright suite the agent then has to maintain. Most "MCP for testing" servers are browser drivers: they hand the agent click/type/screenshot primitives and leave the judging to you.

Prufa's architecture is agentic navigation + deterministic verification: an LLM-backed agent loop plans and drives the browser; a plain-code harness owns the session, records all network traffic over CDP, and grades the captured evidence against a versioned spec. The LLM navigates — it never judges whether a tracking beacon is correct. That separation is why the verdict is trustworthy enough for Kilo Code to act on: same input, same verdict (determinism is regression-tested against a golden fixture site). Prufa's MCP server is a thin client over that — the same product as our CLI, HTTP API and dashboard, just spoken in tool calls. Read the long version in how Prufa verifies a signup flow.

1. 1

   Add the server

   pip install prufa-mcp, then in Kilo Code open Settings → MCP Servers (Agent Behaviour tab) → Add Server and choose Local (stdio) — or edit kilo.jsonc directly under the mcp key. Kilo Code runs as a VS Code extension and a CLI; the same config serves both.

2. 2

   Kilo Code calls a tool

   It runs prufa_run_audit({ url }). Prufa spins up a real browser, walks the public pages, and captures beacons, console output, cookies and response codes — no script for you to write.

3. 3

   Evidence keeps the verdict

   Findings come back in two tiers — verified (a machine-checked fact, in mono) and advisory (a labeled LLM observation, never phrased as broken). Only verified findings move the section grades; there is no composite 0–100 score, ever.

Install, then add the local (stdio) server — the free audit needs no key:

pip install prufa-mcp

// kilo.jsonc  (project ./kilo.json overrides global ~/.config/kilo/kilo.json)
{
  "mcp": {
    "prufa": {
      "type": "local",
      "command": ["prufa-mcp"],
      "enabled": true
    }
  }
}

{
  "mcp": {
    "prufa": {
      "type": "local",
      "command": ["prufa-mcp"],
      "enabled": true,
      "environment": { "PRUFA_API_TOKEN": "{env:PRUFA_API_TOKEN}" }
    }
  }
}

Add the server and let Kilo Code run the first QA test

Or skip the agent and run the free 60-second website audit yourself: paste a URL, get a shareable report in about a minute. Persistent monitors and Gremlin chaos runs start on Pro ($99/mo).