Crux
API Reference@crux/coreIndex Lints

definition.missing_eval_coverage

What it checks

Crux emits this finding when an authored runnable definition is visible in the Project Index but Crux cannot see eval, suite, experiment, baseline, or quality coverage for it.

Why it matters

Eval coverage lets you compare authored AI behavior against known cases and catch regressions before prompts, agents, flows, RAG pipelines, or compositions drift.

How to fix

Add an eval or suite that covers the definition, or connect existing quality records to it so Crux can show affected checks when the definition changes.

When to suppress

Suppress only for prototypes, external test systems, or definitions that are intentionally not regression-tested:

// crux-lint-disable-next-line definition.missing_eval_coverage -- covered by external QA

Rule metadata

  • Rule id: definition.missing_eval_coverage
  • Category: evaluation
  • Maturity: preview
  • Default profiles: recommended, strict
  • Default severity: info

On this page