Turn your traces into verified online evaluators.
AI evaluation
Creates and verifies one LangSmith online evaluator at a time through trace inspection and user-guided iteration.
When to use it
Use when creating online evaluators for LangSmith rather than Harbor-style evaluations.
Give it a LangSmith project name and quality concerns; it proposes options, gets approval, and creates and attaches the selected evaluator.
What you provide
This skill
LangSmith
Creates evaluators and run rules
LangSmith run rule
Enables evaluation of new traces
Requires access to a LangSmith tracing project whose traces can be inspected.
Uses Python to define and run evaluator functions and API workflows.
Requires httpx to create LangSmith run rules through the REST API.