Build AI evaluators grounded in real failures.
What it does
Guides the creation and execution of Phoenix evaluators for AI applications, combining deterministic checks with human-validated LLM judgments.
When to use it
Use when building evaluators or evaluating retrieval and faithfulness in RAG systems.
How to use it
Give it an AI application to evaluate; it builds evaluator code and guides validation against human judgments.
What you provide
- An existing project
Uses
Latest review
What you need · 2
Requires a running Phoenix server.
Remote or cloud Phoenix access requires an API key together with base_url; local access can instead use Client() without explicit credentials.
Access · 1
This skill
Phoenix
Write
Adds review notes to Phoenix