x-bdd-fake-green

A fake green outcome in BDD testing occurs when scenarios pass without verifying real production behavior, often because Given/When/Then steps only mutate local state or inject synthetic results instead of exercising actual code paths.

You detect fake green tests in godog and playwright-bdd by auditing step definitions to ensure they require observable production calls or verifiable external-boundary effects, flagging local error injection patterns and DOM-only assertions that bypass real UI verification.

Handle undefined or pending BDD steps honestly by returning godog.ErrPending when real checks are not currently available, rather than forcing scenarios to green with local state mutations or synthetic success assertions.

BDD test integrity auditing flags local error substitutions like LastErr assignments or synthetic gRPC status injections, identifying success-by-any-error assertions that mask failures and break honest pass/fail semantics.

Your BDD scenario passes when the service contract is broken because the step definitions manipulate local state without making observable production calls, creating a fake green outcome that silently masks missing verification.

BDD fake green detection requires step definitions to be auditable for observable production calls or external-boundary effects, and cannot verify scenarios where production paths are inherently inaccessible or where local state mutations are indistinguishable from real behavior.