Create a canonical local job record, capture revisions and invalidate dependent checks when critical assumptions change.
A 500-copy brochure is approved at A4 on coated stock. After approval, quantity rises to 1,500. The artwork file may be unchanged, but capacity, material demand, packaging and cost are no longer the same decision. Saving the quantity change as a new revision preserves the original baseline and marks the affected downstream checks for review.
A later trim-size change is more disruptive. It can alter crop, effective PPI, imposition, sheet yield and provider capability at the same time. The core does not recalculate those specialist results itself; it records which evidence became stale and routes the operator back to the right modules.
Use the revision note to explain why a change happened, not merely what field changed. A short reason such as ‘client reduced finished width after proof’ gives future reviewers context that an automatically generated diff cannot provide.
Changing quantity, trim, source files or deadline can invalidate downstream answers. The core stores explicit revision snapshots so a later calculation is not silently treated as current.
The engine marks affected areas for re-check; it does not pretend to know a provider-specific consequence it cannot prove.
Every record is JSON-compatible and can be exported through the Vault instead of being trapped in a hidden cloud account.