Automate eval-driven development workflows for Claude Code sessions.
npx skills add https://github.com/ldk-hub/broke-shopping --skill eval-harness-ldk-hub
Please help me install this Agent Skill. Skill: eval-harness Source: https://github.com/ldk-hub/broke-shopping/tree/main/.agent/.agents/skills/eval-harness Command: npx skills add https://github.com/ldk-hub/broke-shopping --skill eval-harness-ldk-hub
Formal evaluation framework for Claude Code sessions enabling eval-driven development (EDD) principles.
Define your first evaluation and run it against Claude Code to generate a results report.
High-intent search queries and answers about installing and using this skill.
Regression evaluations for Claude Code are run by defining capability checks, executing deterministic and model-based graders, and generating results reports stored under .claude/evals.
Eval-driven development for model evaluation is a formal workflow that applies deterministic and model-based graders to assess capability and regression across prompt iterations and model versions.
Yes, pass@k metrics are supported within the evaluation framework to measure capability and regression outcomes across different model versions and prompt iterations during Claude Code sessions.
Define evaluation definitions for Claude Code by specifying capability and regression checks, which are then executed by deterministic and model-based graders to produce stored evaluation data.
Yes, the evaluation framework stores all evaluation data and results locally under the .claude/evals directory, ensuring persistent access to capability and regression test outcomes.
Model-based graders in regression evals may exhibit non-deterministic scoring behavior across model versions, requiring deterministic checks to validate baseline capabilities and ensure stable prompt iterations.
zhmxiaowoDefine automated evaluation criteria for Claude Code tasks with multiple graders.
xu-xiangDefine, execute, and report capability and regression evals for Claude Code sessions.
T-ravAutomate Claude Code session evaluations with capability and regression tests.
choub00Evaluate Claude Code sessions with capability and regression evals.
