Define and run capability evals with code and model graders.
npx skills add https://github.com/fancive/claude-skills --skill eval-fancive
Please help me install this Agent Skill. Skill: eval Source: https://github.com/fancive/claude-skills/tree/main/skills/eval Command: npx skills add https://github.com/fancive/claude-skills --skill eval-fancive
Eval-Driven Development enables teams to define and manage capability/regression evaluations before coding, ensuring measurable quality throughout the lifecycle.
Start by creating an eval definition with /eval define <name>, then run checks with /eval check <name> and finally generate a report with /eval report <name>.
High-intent search queries and answers about installing and using this skill.
Yes, you can apply eval definitions across planning, development, and post-implementation review. The workflow supports establishing benchmarks early, validating changes during development, and generating final reports to ensure measurable capability and regression checks are met.
Eval-driven development is the process of defining capability and regression evaluations before writing code. You need it when you want to establish measurable quality benchmarks, ensuring that development progress is validated against baseline metrics rather than subjective standards.
To run code-based and model-based graders, you execute a check command against your defined eval name. This validates your changes by running the graders and automatically saving the results and baselines for comparison against future runs.
You generate a report with pass@k metrics by executing a report command with your eval name. This produces a comprehensive report summarizing the evaluation results, allowing you to review capability scores and regression status.
Yes, you can apply eval definitions across planning, development, and post-implementation review. The workflow supports establishing benchmarks early, validating changes during development, and generating final reports to ensure measurable capability and regression checks are met.
ldk-hubAutomate eval-driven development workflows for Claude Code sessions.
GoBeromsuDefine and run evaluations for Claude Code changes with structured graders and reports.
JamkrisAutomate eval-driven testing workflows for Gemini CLI sessions with deterministic graders.
hieuckCoordinate eval-driven development workflows for Claude Code sessions.
