skills/smithery.ai/build-eval

build-eval

Installation
SKILL.md

Eval-1337

Write evals that measure what matters. Not vanity metrics.

The key to success is measuring performance and iterating (Anthropic 2026).

The Core Problem

# This tells you NOTHING
activation_rate = 100%  # Activates on every prompt = useless

Single metrics lie. You need to measure BOTH failure modes.

Three Grader Types

Installs
1
First Seen
Apr 22, 2026
build-eval from smithery.ai