Test Quality
Coverage tells you what runs. Not what it catches.
80% coverage means nothing if assertions are shallow and mocks replace every real dependency. Octokraft measures test effectiveness: assertion quality, structural coverage, and whether your tests actually verify behavior.
Four metrics that tell you if your tests actually work
Coverage is a starting point, not a finish line. These four dimensions tell you whether your test suite catches real bugs or just exists to hit a number.
Structural Coverage
Which code paths are actually exercised, not just imported
Assertion Density
How many meaningful checks per test, not just 'expect(true).toBe(true)'
Test/Code Ratio
The balance between production code and the tests that guard it
Mock Usage
Flags tests that mock so aggressively they're only testing the mocks themselves
Surface the test smells that silently weaken your suite
Shallow assertions, repetitive templates, existence-only checks. These pass CI every time but catch nothing when behavior changes. Octokraft flags them with severity, file path, and confidence so you can fix the ones that matter.
- Shallow assertions that verify a function was called but not what it returned or how it behaved.
- AI-generated test smells: repetitive template structures with minimal variation that add coverage without value.
- Existence-only assertions on static metadata that will never catch a regression.
- Every finding includes the file path, category, and confidence score so you can triage fast.
Track whether your test suite is getting stronger or weaker
A one-time audit tells you where things stand. Tracking over time tells you if they're improving. Every health assessment recalculates test quality so you can see the trend.
- Test quality scores update on every merged PR. You'll know immediately if a change weakened your suite.
- Assertion density trends reveal whether new tests are meaningful or just padding coverage.
- Mock usage trends show if the team is testing real behavior or drifting toward isolated mocks.
Find out if your tests actually catch bugs
Not just coverage. Assertion quality, mock patterns, and structural depth. Measured on every commit.
Try Octokraft