Create Developer Documentation for Evaluation Features

[maxtechera]

Objective

Create developer documentation for extending and integrating with evaluation features.

Tasks

API Documentation

• Document all dataset API endpoints with examples
• Document all evaluator API endpoints with examples
• Document evaluation execution endpoints
• Document request/response schemas
• Add curl examples for common operations

Architecture Documentation

• Document entity relationships (ERD)
• Document service layer architecture
• Document evaluation execution flow
• Document how simple evaluators work
• Document how LLM evaluators work

Extension Guide

• Document how to add custom simple evaluators
• Document how to add custom LLM evaluator templates
• Document integration points for external evaluation systems
• Document how to extend evaluation metrics

Integration Examples

• Show programmatic dataset creation
• Show running evaluations via API
• Show retrieving results programmatically
• Show webhook integration (if supported)

Location

• packages/server/CLAUDE.md - Add evaluations section
• packages/components/CLAUDE.md - Add evaluator components section (if applicable)

Acceptance Criteria

• All API endpoints documented with examples
• Architecture clearly explained with diagrams
• Extension guide enables adding custom evaluators
• Code examples are tested and working
• Documentation follows existing CLAUDE.md patterns

[linear]

AGENT-605