It reads your code, finds what will break under load, builds the test, runs it on real infrastructure, and tells you exactly why it failed — end to end, no dashboards, no YAML, no guesswork. Most load-testing tools hand you an empty script editor and a shrug. This is the opposite. Drop the skill into your AI agent and just say "help me load-test this." It runs a real engineer's playbook: Every res
npx mdskills install MaxoPerf/agent-skillComprehensive end-to-end load testing orchestration with detailed workflow and strong guardrails
It reads your code, finds what will break under load, builds the test, runs it on real infrastructure, and tells you exactly why it failed — end to end, no dashboards, no YAML, no guesswork.
npx @maxoperf/agent-skill install
Most load-testing tools hand you an empty script editor and a shrug. This is the opposite. Drop the skill into your AI agent and just say "help me load-test this." It runs a real engineer's playbook:
| Step | What it does |
|---|---|
| 🎯 Intake | Asks one sharp question at a time until it knows your goal — baseline, breaking point, soak, or "just show me the numbers." No jargon; you never have to speak in "virtual users." |
| 🔍 Hotspot hunt | Opt-in: scans your repo for the endpoints that will fold first — N+1 queries, unbounded reads, fan-out calls, blocking work on the request path — and ranks them by blast radius. Framed as opportunities, never criticism. |
| 🧠 Engine choice | Picks the right load generator for the job — k6, JMeter, Playwright, Selenium, Gatling, Locust — with a one-line rationale you can override. |
| ✍️ Builds the test | Writes the actual script into your repo, with your pass/fail SLOs baked in as thresholds. Secrets go to the vault, never inline. |
| 🚀 Runs it for real | Uploads and fires the run on MaxoPerf's managed cloud runners — hundreds of thousands of VUs if you want them — and streams you the console link so you can watch live. |
| 🕵️ Root-cause analysis | When something breaks, it doesn't just say "failed." It correlates error bodies, status codes, engine logs, and runner health to tell you why — app error vs. failed SLO vs. under-provisioned load vs. a flaky dependency. |
| 📈 Anomaly detection | A deterministic outlier scan (robust median/MAD, tuned for near-zero false positives) flags the latency spikes and throughput cliffs you'd have missed — and it knows to run after the test ends, when the data is real. |
Every result comes with the numbers and a one-click breakout to the full MaxoPerf console. You're never locked in the chat.
vuShortfallPct before it ever blames your app.Any AI agent (Claude Code, Cursor, Codex):
npx @maxoperf/agent-skill install
Prefer the one-click bundle? Install it together with the MCP server via the Claude Code plugin:
/plugin marketplace add MaxoPerf/mcp
/plugin install maxoperf
You'll need a MaxoPerf API key (console → Settings → API keys) — the skill will walk you through wiring it the first time.
The skill is the brain; the MaxoPerf MCP server is the hands. The skill does the thinking — reading your code, choosing engines, writing tests, diagnosing failures — and drives the hosted MCP server's tools to actually create, upload, run, and analyze on MaxoPerf's platform. One connects to the other automatically.
Get started → · MCP server → · Docs →
Read-only mirror of the MaxoPerf monorepo. File issues at maxoperf.com.
Install via CLI
npx mdskills install MaxoPerf/agent-skillMaxoPerf Agent Skill is a free, open-source AI agent skill. It reads your code, finds what will break under load, builds the test, runs it on real infrastructure, and tells you exactly why it failed — end to end, no dashboards, no YAML, no guesswork. Most load-testing tools hand you an empty script editor and a shrug. This is the opposite. Drop the skill into your AI agent and just say "help me load-test this." It runs a real engineer's playbook: Every res
Install MaxoPerf Agent Skill with a single command:
npx mdskills install MaxoPerf/agent-skillThis downloads the skill files into your project and your AI agent picks them up automatically.
MaxoPerf Agent Skill works with Claude Code, Claude Desktop, Cursor, Vscode Copilot, Windsurf, Continue Dev, Codex, Gemini Cli, Amp, Roo Code, Goose, Opencode, Trae, Qodo, Command Code, Chatgpt. Skills use the open SKILL.md format which is compatible with any AI coding agent that reads markdown instructions.