Skill UI

This skill executes a corpus benchmark to verify the performance and cost claims of the cost booster system. It runs local booster tests against optional baselines like Gemini or Anthropic models. Results are persisted to JSON files for verification gates before releases. It checks win rates and escalation rates to ensure quality improvements and validate upstream tags.

Master comprehensive evaluation strategies for LLM applications, covering automated metrics like BLEU and ROUGE, human evaluation dimensions, and LLM-as-Judge patterns. This skill helps measure performance, compare models, detect regressions, and validate prompt improvements. Includes Python implementation examples for metrics and evaluation suites.