aibdd-green-execute

To make failing BDD acceptance tests pass, you can use product-code-only repair loops that edit application logic iteratively until the target feature files turn green, leaving BDD specs and DSL artifacts untouched.

Runtime drift detection in acceptance testing verifies whether the execution environment or test visibility has changed between handoffs. You need it before editing product code to ensure failure evidence remains valid and actionable.

To fix oscillating acceptance test failures, apply oscillation detection guardrails that identify looping fixes and emit a structured stop reason, routing execution back to red when evidence becomes invalid or non-actionable.

No, automated BDD green execution requires a completed legal red handoff as input. It enforces handoff validity checks upfront and stops if the red state is already green or contains invalid failure evidence.

The minimal edit strategy for acceptance-test-driven repair repeatedly edits only the smallest amount of product code needed to clear the first failing scenario, iterating until the entire target feature set passes.

Acceptance test repair stops early when it detects oscillation, runtime drift, architecture vetoes, or non-actionable failures, emitting a precise stop_reason for routing rather than attempting blind fixes that could destabilize the codebase.