stage-contract-change

To update stage contracts safely, you must coordinate changes across contract documents, validator logic, runtime prompts, and test scenarios simultaneously. This prevents mismatched validators and inconsistent outputs that halt AI delivery progress.

Code-only contract changes are blocked by safety guardrails. You must update contract documents, validator logic, and test scenarios together to prevent implicit output widening and ensure workflow validation remains functional.

Adding a required field to a stage output contract requires coordinated updates to all associated validators, runtime prompt instructions, and test scenarios. This ensures the workflow processes the new field without generating invalid stage outputs.

Workflow validation fails after modifying stage input documents because piecemeal changes cause mismatched validators and inconsistent prompts. You must synchronize contract updates with validator logic and repair behavior to restore functionality.

No, non-Markdown canonical output forms are strictly prohibited by hard rules. The stage contract update process mandates Markdown to maintain consistency and enforce safety guardrails across governed AI software delivery workflows.

You must synchronize validators immediately after changing runtime prompt instructions to prevent invalid stage outputs. Coordinated updates ensure validation rules match the new prompt behavior and accompanying test scenarios verify the changes.