evals-create-suite

Scaffold LLM evaluation suite packages with Playwright configuration files.

21.2k|8.6k|Updated Jan 26, 2013
One-click install
npx skills add https://github.com/elastic/kibana --skill evals-create-suite
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: evals-create-suite
Source: https://github.com/elastic/kibana/tree/main/.agents/skills/evals-create-suite
Command: npx skills add https://github.com/elastic/kibana --skill evals-create-suite

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill automates the setup of new LLM evaluation suites, ensuring consistent project structure and boilerplate code.

Core Features & Use Cases

  • Scaffolds new eval suite packages: Creates directories, configuration files (playwright.config.ts, package.json, kibana.jsonc, tsconfig.json), and basic evaluate.ts re-exports.
  • Ensures correct Playwright configuration: Places playwright.config.ts in the package root, avoiding Scout CI discovery issues.
  • Use Case: When starting a new LLM evaluation project for a specific feature or plugin, use this Skill to quickly generate the foundational files and directory structure.

Quick Start

Use the evals-create-suite skill to scaffold a new eval suite named 'my-new-eval' in the 'x-pack/solutions/ai' directory.

Frequently Asked Questions about evals-create-suite

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I scaffold a new LLM evaluation suite with Playwright?

Scaffolding a new LLM evaluation suite generates directories, configuration files like playwright.config.ts, package.json, and evaluation fixtures to provide consistent boilerplate for offline LLM evaluations.

What is the best way to set up boilerplate for offline LLM evaluations?

The best way to set up boilerplate for offline LLM evaluations is using an automated scaffolding process that creates package files and evaluation fixtures, ensuring correct project structure and consistent configuration across new eval suites.

Do I need specific inputs to create an eval package for plugins?

Yes, creating an eval package for plugins requires specific inputs including the suite name, parent directory, owner, group, visibility settings, and any custom fixture needs to correctly configure the generated package files.

Why does my Playwright configuration need to be in the package root for Scout CI?

Playwright configuration must be in the package root to avoid Scout CI discovery issues, ensuring the evaluation suite is correctly identified and executed during continuous integration testing workflows.

Can I add custom evaluation fixtures when creating an LLM eval suite?

Yes, you can specify custom fixture needs when creating an LLM eval suite, allowing the scaffolding process to generate tailored evaluation fixtures alongside the standard package configuration files.