Parallel Test Bench

Infrastructure for running parallel agent tests with automatic metric collection, programmatic grading, and benchmark aggregation. Use this skill when the user wants to A/B test different approaches, compare agent configurations, benchmark performance improvements, run parallel experiments, or test multiple variations of a prompt or implementation. Trigger for phrases like "test in parallel", "compare approaches", "benchmark", "A/B test", "run experiments", or when the user wants to collect timing/token metrics from multiple agent runs.

89jobrien Updated 1 repo stars

File contents

89jobrien/pjlib commit 0264cd2c6c

Frequently asked questions

npx get-skill add 89jobrien/parallel-test-bench