Design a regression test for a fix in Claude Code

Capture observable failure before a fix, keep the test after recovery, and size coverage to module stability.

Install in Claude Code

A Claude Code project and a working test framework. This example uses a pure function without network access.

Review `.claude/skills/testing/SKILL.md` and its requirements at the pinned GitHub commit.
Create `.claude/skills/testing/` in your project and save the reviewed file as `.claude/skills/testing/SKILL.md`.
Start Claude Code from that project's root, then select `/testing` in the slash menu.

Example task

An event-filter function throws on empty input. The contract says `{events:[],venue:'Hall'}` should yield []. This input reproduces the incorrect behavior before the fix.

What to check

Write one isolated regression test for `Array.isArray(result) && result.length === 0`. Run the same assertion after the fix and report the real outcome. Keep the existing normal-filter test.

Original SKILL.md on GitHub · Updated 2026-11-14

Frequently asked question

Why keep the same assertion when fixing empty-input behavior?

Reproduce the bug reliably and assert the public result. Use the same isolated test to verify the fixed behavior, independent of implementation details.