pivot-or-perseverelisted

# Pivot or Persevere The most expensive decision in a startup is usually not deciding. This skill turns the pivot-or-persevere question into a scheduled, evidence-backed meeting with a written verdict — and makes "give it another quarter" an explicit, threshold-bound commitment instead of a flake. ## The Rule Persevering in a direction with no validated learning is not grit; it is a slow-motion pivot you have not admitted to yet. Conversely, pivoting in reaction to a single bad week is panic dressed as decisiveness. The verdict must rest on cohort evidence over a defined window — never on the last board meeting's mood. ## Modes ### Prep mode (user faces the decision but has not assembled evidence) 1. Read `references/pivot-decision-rules.md` before structuring anything. 2. Establish the evidence base: baseline metrics at start, cohort trend since, and the learning milestones actually completed. If the team cannot state its riskiest assumption and whether the last loop validated it, say so — that is the first finding, and it usually means the problem is measurement, not direction. 3. List the failure conditions from the reference and check each against the evidence. 4. Output the meeting agenda from `templates/pivot-or-persevere-meeting.md`, filled with the evidence available and marked UNKNOWN where evidence is missing. ### Run mode (user is in or about to enter the meeting) 1. Walk the agenda in order. Do not let the meeting skip the evidence section and