Eval Harness Skill

A formal evaluation framework for Claude Code sessions, implementing eval-driven development (EDD) principles.

When to Activate

Setting up eval-driven development (EDD) for AI-assisted workflows
Defining pass/fail criteria for Claude Code task completion
Measuring agent reliability with pass@k metrics
Creating regression test suites for prompt or agent changes
Benchmarking agent performance across model versions
Designing a shadow evaluation pipeline to compare a new model against the production champion without user exposure
Implementing an LLM-as-judge grader for open-ended outputs where deterministic grading is not feasible
Setting up a delayed-label ground-truth pipeline for production ML systems