vector-forge

Mutation testing for cryptographic implementations finds surviving mutants by altering code and checking if baseline tests fail. It identifies uncovered paths and missing assertions, revealing exactly where your test suite lacks the vectors needed to catch deep arithmetic or rejection issues.

Call-graph reachability triages escaped mutants by mapping which code paths are actually exercised by your tests. It classifies escapes into actionable targets like missing vectors, fuzzing targets, or false positives, ensuring you only generate vectors for genuinely uncovered code.

You create negative test vectors by crafting Wycheproof-style inputs that deliberately exercise invalid code paths and assert rejection. Including roundtrip checks ensures you catch deep arithmetic issues alongside the rejection logic for robust cryptographic testing.

Yes, you measure before and after kill rates by comparing mutation survival rates before adding new vectors and after. This comparison directly proves vector value by showing how many previously escaped mutants the new Wycheproof-style vectors successfully kill.

The best way to close cross-implementation coverage gaps is to use mutation testing to discover escaped mutants, generate Wycheproof-style vectors targeting those paths, and perform cross-implementation verification to ensure the vectors catch regressions across different suites.

Yes, a baseline mutation run is a required workflow step. You must first run mutation testing on your cryptographic implementation to discover which mutants escape, then use those escaped mutants to inform the generation of targeted test vectors.