How can you estimate a task when important requirements are still unclear?

Separate known work from assumptions and identify the uncertainties most likely to change the estimate. Ask targeted questions or time-box a small investigation, then provide a range with its assumptions rather than false precision. Include integration, review, and verification, not only implementation time. Break the work into smaller outcomes and revise the estimate as uncertainty falls. Explain what would trigger a change in scope or schedule. Estimates help planning; they should not hide missing information by converting a guess into an unconditional commitment.