Focus

Turn a product opportunity into a brief the human can decide on. The human owns product priorities and business rules; label proposals and assumptions separately from confirmed decisions.

1. Identify who is affected, what they are trying to do, and the consequences of the current experience. Separate observed behavior, reported needs, and hypotheses. Research outside evidence when it would change the decision; use Curate for sourced comparisons of options.
2. Define the desired outcome, constraints, goals and non-goals. Compare viable choices, including keeping the current behavior when plausible, using user impact, cost, and uncertainty. For each proposed success measure, say how it could be observed and mark an unknown baseline or unapproved target as such.
3. Recommend a priority with the evidence and tradeoffs behind it. For an uncertain product requirement or business rule that affects the outcome, present the evidence and options and ask the human to decide. State which decisions still need an answer. If an authorized experiment would resolve a concrete product uncertainty, use the prototype method to test scenarios without treating simulated behavior as approved policy. For example, a refund model can expose what happens after a partial return; the refund rule remains the human's decision.

Done: Deliver the product brief: user problem and evidence, goals and non-goals, options, success measures, recommendation, and decisions needed. If saving a standalone brief without a designated destination, use .work/briefs/ under the workspace rules. Use Plan to sequence agreed work when requested; use Brief for instructions to transfer work to another worker.