Build LangSmith evaluators from real agent outputs.
AI evaluation
Builds LangSmith evaluation pipelines with evaluators, run functions, and local or uploaded evaluation runs.
When to use it
Use when building evaluation pipelines for LangSmith.
Give it an agent project; it creates evaluators and run functions for local or uploaded evaluation.
What you provide
This skill
LangSmith evaluator rules
Changes LangSmith evaluator rules (irreversible)
LangSmith Cloud feedback
Sends feedback to LangSmith Cloud (irreversible)
LangSmith
Reads your LangSmith data
LangSmith evaluations
LangSmith evaluation jobs
Requires access to LangSmith through a saved CLI profile or API key.
An OpenAI account is optional and enables locally executed OpenAI judges.