A modal verb is a useful clue but an unreliable approval signal
The appendix label changes the evidential role of the sentence even though its grammar remains imperative. A reviewer should see the training context before seeing any proposed date field. Otherwise a scheduler may manufacture an urgency that no customer requested, and later participants may mistake that date for an agreement already checked by someone else. One alternative is to exclude all examples during extraction, but that would also hide examples used to clarify genuine obligations. A better review record keeps the sentence, its classification, and the reason it was excluded from committed scope. The contrasting test should use identical wording in different document roles. Success means the reviewer can explain why one occurrence supports a deadline and the other supports only an illustration, without relying on a hidden classifier score.
Working through the example
A training appendix says that a sample supplier shall deliver on Monday. The actual order has no delivery date, yet an extractor proposes Monday as a project deadline. Keep the appendix's example label with the excerpt. Classify the sentence as illustrative material and ask the project owner for the authoritative delivery condition.
A discriminating check
Test with the same sentence moved between an examples appendix and a signed scope section. Document headings may themselves be ambiguous; human classification still needs the surrounding document and its purpose. The project’s DemaRQ research note motivates distinguishing requirement text from examples and explanation. Its sentence-classification setting does not establish approval or resolve the surrounding business context.