sprint_preplisted

# sprint_prep — declare the riverbed Load `sprint_protocol` first. Use this as the owning Planner while a Sprint is `prepared`. Preparation ends at one atomic arming decision; it does not launch participants piecemeal. ## Outcome Produce one editable prepared Sprint with: - one roadmap feature; - exact governing spec revision hashes and any optional QA/QC evidence; - work units made from existing spec tasks, each with one Developer and one assigned Reviewer; - dependency edges and planned waves; - one harness/model/Thinking level (`effort`) intent per participant; - the Sprint merge grant — the FnB's merge authorization for every registered Sprint PR, given by deciding to run the Sprint and recorded with `--merge-grant`; the engine refuses to declare or arm without it; and - a capacity plan sized to justified parallel work and review demand, with the local/GitHub capacity to execute it. Preview every participant with `sc models resolve <harness> [<model>] [--effort <level>]`; omit `--effort` for Vibe and Harness default. Pass = each controlled preview names the requested Thinking level and each uncontrolled preview returns explicit `effort: null`. Arm binds every route, records the armed transition, and publishes the first assignments in one transaction; any mismatch rolls the whole arm back. ## Eligibility pass Read the feature, selected spec bodies, task ledgers, available QA/QC records, shell roster, model routes, quota state, repository access, and worktree