adversarial-verify

Adversarial verification probes target files and change sets for robustness under malformed inputs, concurrency hazards, and invalid states. You provide explicit target files, a change description, and a risk tier to execute structured probes against hostile conditions.

Adversarial testing applies tier-aware attack vectors to implementation code and configuration updates to identify security defects and verify rollback behavior. It records executed probes and reproduction details in a structured report with a final verdict and evidence.

You verify recovery by running adversarial probes against configuration updates to test failure paths, state integrity, and rollback behavior. The Skill requires a change description and target files to execute probes that document how the update handles partial failures.

Adversarial code review works with implementation code, configuration updates, contracts, and change sets. It applies tier-aware attack selection to identify defects under malformed inputs, invalid states, concurrency, and recovery scenarios across these target file types.

You need to provide explicit target files, a change description, and a risk tier. The verification process requires these inputs to select appropriate attack vectors, execute probes, and generate a final report with a verdict and reproduction evidence.

You should not use adversarial probing when you lack explicit target files or a defined change description. The verification process depends on these inputs to select tier-aware attack vectors and execute probes that produce a structured report with a verdict and evidence.