/demo

The heatmap, live.

This is a real StrengthReport — the JSON speccle-oracle strength --json emits — produced from the repo's targets/checkoutproving ground. Line coverage is a perfect 100%, and two mutants still survived. Click a weak criterion to see the exact changes its tests didn't notice, or flip to the coverage view to see what coverage alone would claim.

Line coverage

the naïve baseline

Oracle strength

killed ÷ covered mutants

Gap

4.3%

what coverage overstates

46 covered mutants: 44 killed, 2 survived — each survivor a change no test noticed

A criterion is strong only when every mutant its tests cover is killed.

features/basket/SPEC.md

strength 100.0% · coverage 100.0%

BASKET-1100%Adding an item increments its quantity by exactly 116/16 killed

BASKET-2100%Removing the last item leaves the basket empty29/29 killed

features/checkout/SPEC.md

strength 88.2% · coverage 100.0%

CHECKOUT-1100%Tax rounds half-up to 2dp per line item14/14 killed

CHECKOUT-2100%An empty basket totals zero7/7 killed

Nothing falls outside the criteria here: every criterion is claimed by a test, no test claims an unknown token, and no mutant sits beyond a claiming test's reach.

The four states

This example happens to exercise the first two; a criterion can also be unclaimed or uncovered.

Strong

Every mutant this criterion's tests execute is killed — the behaviour is genuinely defended.

Weak

At least one surviving mutant — a code change the tests executed but didn't notice. This is what strengthen routes.

Unclaimed

No test carries this criterion's [KEY-n] token. Reported by name as the worst case — never as a zero score.

Uncovered

Tests claim the criterion but cover no mutants — they execute nothing the mutation run touched.