Write a plugin development guide with step-by-step example

[himanshu231204]

Description

OpenAgent Eval supports plugins via Python entry points, but there's no dedicated developer guide showing how to create a custom metric, provider, or dataset loader from scratch. The existing \plugins/examples/custom_metric.py\ is a good starting point but needs documentation to go with it.

What to Add

Create a new page \docs/plugin-development.md\ (or add to the existing plugin docs) covering:

1. Overview of the plugin system architecture
2. Step-by-step: Creating a custom metric (using WordCountMetric as example)
3. Step-by-step: Registering a custom LLM provider
4. How to configure \pyproject.toml\ entry points
5. How to test a plugin locally
6. Best practices and common pitfalls

Files to Look At

• \openagent_eval/plugins/\ — plugin system implementation
• \plugins/examples/custom_metric.py\ — example plugin
• \pyproject.toml\ — entry point configuration
• \docs/\ for MkDocs integration

Good First Issue ✅

A documentation task that helps you understand the plugin architecture deeply.