/demo

The heatmap, live.

This is a real ReportModel — the JSON the speccleCLI emits, produced from the repo's own test fixtures for a small math module. Click a criterion to inspect its tests and mutants; drag θ to see how the strong/weak line is drawn.

Line coverage

the naïve baseline

Oracle strength

killed / total mutants

Gap

12.5%

what coverage overstates

8 mutants: 5 killed, 1 survived, 2 uncovered

Threshold θ 0.80

Oracle view: a criterion is strong only when its linked tests kill mutants at or above θ.

Arithmetic

Unimplemented

AC-3—pow raises the base to the exponent.no tagged tests

Unattributed mutants 4reached by no token-tagged test — they still count against oracle strength

src/math.ts:13BlockStatement{ return a / b; }→{}survived

src/math.ts:14ArithmeticOperatora / b→a * bsurvived

src/math.ts:9BlockStatement{ return a * b; }→{}killed

src/math.ts:10ArithmeticOperatora * b→a / bkilled

The four states

This fixture happens to exercise three of them; a criterion can also be inert.

Strong

Linked tests kill mutants at or above the threshold — the criterion is genuinely defended.

Weak

Tests cover the code but mutants survive — often a spec problem, not just a test gap.

Unverified

No test carries this criterion's [AC-n] token.

Inert

Linked tests exist but cover no mutants — they touch nothing the criterion defends.