MaxoPerf agents — build and run tests from any AI coding agent

An AI agent can drive MaxoPerf as well as a person can. Any agent that speaks MCP can call MaxoPerf’s tools directly. The MaxoPerf skill adds the workflow. It is a small, shippable instruction bundle for skill-aware agents (Claude Code, Claude Desktop, and compatible tools). It tells the agent how to pick a test type, where secrets go and what a healthy run looks like. Without it, you would repeat that in a long prompt every session.

Get started with the MaxoPerf skill.

What the skill teaches an agent

The skill takes you from “I have an app” to an uploaded, running MaxoPerf test in a fixed sequence. Every step uses shipped platform facts, never invented ones:

1. Intake, one question at a time. The agent asks what to test, where the target is, roughly how much load to apply and what counts as pass/fail. If the target is local or not deployed, the agent offers a reverse tunnel.
2. An optional hotspot scan. If you ask, the agent greps your repo for likely load-testing hotspots (heavy routes, N+1 queries, fan-out calls) and offers a short list. The scan is read-only, and nothing leaves the agent’s own context.
3. Choosing an executor from the platform’s supported-executor list: k6 or jmeter for API/protocol targets, playwright or selenium for browser journeys. If you already have an asset (a JMX file, a Locust script, a .side recording), the agent reuses it. See Tests & builders for everything MaxoPerf can run.
4. Building the test artifact in a shape the platform accepts. For browser tests the skill knows every supported script shape: Playwright spec/library/IIFE, Selenium test-function/bare script, or a scriptless interactive journey. So the agent does not rewrite your existing script when it already fits.
5. Uploading and running it with the MCP tools create_test, upload_test_file and start_run, then polling until the run reaches a terminal state.
6. Reading results back, with a console deep link. On a failed or suspicious run, the agent finds the root cause with the diagnostics tools (get_run_errors, get_run_logs, get_runner_health, detect_run_anomalies).
7. API-key onboarding the first time the agent hits an auth error. A fresh session gets connected without leaving the conversation.

The same skill covers multiple environments (staging vs. production, or per-region targets). It also covers features beyond the core loop: tunnels, browser fleets, virtual services and on-demand browser inspection for single-page apps. The agent uses them only when the situation needs them, not by default.

Zero invented facts, by design

Every executor name and tool argument in the skill is pinned to MaxoPerf’s shipped source. The skill’s reference pages are the same source Loadrigo and this documentation site use. If a capability isn’t real, the skill doesn’t claim it.

Install the skill

1. Get a MaxoPerf API key first. See Get an API key. Keep it at hand, because the installer asks for it.

2. Run the installer from a terminal where your agent (Claude Code, etc.) is available:

   npx @maxoperf/agent-skill install

3. Paste your API key when prompted. The installer stores it the same way a manual MCP config would. It never prints or logs the key.

4. Restart or reload your agent so it picks up the new skill and MCP connector.

The installer is idempotent. Re-run it any time to update the skill in place. For the full walkthrough, including uninstall and first-run prompts to try, see Install the MaxoPerf skill.

No skill support? Connect over MCP directly

If your agent doesn’t support skills, it can still do everything above by calling MaxoPerf’s tools once connected. See Connect your AI agent to MCP for per-client setup and Tools for the full catalog. Chat-only clients that can’t read a filesystem (Claude Desktop, ChatGPT) get the same intake → build → run flow through the server’s guided prompts.

To connect by hand instead of installing the skill, open Settings → Connect AI tools. It shows the endpoint and a config block for each client, ready to paste.

Safety

An API key grants full-account access. That holds for a key in a manual MCP config and for one the skill installer stores. Review your agent’s tool-call log before you approve create/run/delete actions, especially the first few times. Treat the connection like any other credential. See Get an API key for scoping and rotation guidance, and read-only mode if you want an agent to investigate without being able to change anything.

Where to go next

MCP overview The endpoint, authentication, and read-only mode underneath every agent connection.

MCP tools reference Every tool, prompt, and resource the skill and any other agent can call.

Loadrigo MaxoPerf's own in-console assistant, built on the same tools and skill.

Tests & builders What kinds of tests an agent (or you) can build: performance, functional API, browser, scriptless.