What problem does it solve? Broad security and robustness reviews tend to either miss real issues outside a narrow scope or flood you with unverified findings. This Skill provides a structured audit process that achieves both coverage and accuracy by separating hypothesis generation from independent falsification. ## Core Features & Use Cases - Entry-point grid mapping: Builds a matrix of input entry points (URLs, config files, API/CLI arguments, environment variables, persistent storage) crossed with failure classes, so no cell is left unexamined. - Independent falsification phase: Sends hypotheses to a separate refutation lane that tries to kill each claim via upstream rejection, downstream absorption, false premises, or existing tests, keeping only verified findings. - Test promotion: Promotes surviving invariants into unit tests that measure properties across the whole catalog, preventing regressions of the same failure class elsewhere. - Use Case: When asked to comprehensively audit what an attacker or malformed input can do through every external entry point of an application, run this process to produce a verified findings list with reproduction steps and regression tests. ## Quick Start Audit this application's external input entry points for robustness using the adversarial audit process and report only falsification-surviving findings.
