What problem does it solve? Existing cryptographic test vector suites often leave implementation code paths unexercised, and there is no systematic way to measure or close those coverage gaps. This Skill uses mutation testing to find escaped mutants, then generates new test vectors that deliberately target the uncovered paths, proving effectiveness with before/after kill-rate comparisons. ## Core Features & Use Cases - Mutation-Driven Gap Analysis: Runs mutation testing frameworks (gremlins, cargo-mutants, mutmut, Mull, Stryker) against implementations to establish a baseline and identify survived or uncovered mutants. - Graph-Informed Triage: Uses Trailmark call graphs to classify escaped mutants by reachability, blast radius, and security impact, filtering out false positives like equivalent mutations and cross-package coverage artifacts. - Targeted Vector Generation: Produces Wycheproof-format JSON vectors including single-fault negative vectors, roundtrip assertions, and fault-simulation vectors from limb-width reimplementations, cross-verified against two or more independent implementations. - Use Case: When auditing a BLS12-381 library, run the workflow to discover that flag-validation mutations survive, then generate negative deserialization vectors that kill those mutants and document the kill-rate delta in a coverage report. ## Quick Start Use the vector-forge skill to run mutation testing on my BLS12-381 implementations and generate new Wycheproof test vectors that close the escaped-mutant coverage gaps.
