setuplisted

Initialize (or reconfigure) ultragoal for this project. Takes about a minute. ## 1. Knobs Ask the user the seven questions below in **one** AskUserQuestion call (recommended default first). If `.ultragoal/config.md` already exists, show current values and ask only what they want changed. If running non-interactively, take all defaults. 1. **Rigor** — how much scaffolding the harness wraps around the model; match it to model strength. **vanilla** (default — for strong models like Fable: the lean single-grader loop, no extras) / **standard** (bells & whistles: interim re-checks, pessimistic double-runs, research scouts, the background log monitor) / **max** (every recommended technique, for lower-intelligence models or release-grade stakes: 3-lens panel verification, every-claim cadence, multi-modal scout sweeps, deep interview, rubric variants). The goal skill reads this and shapes the whole loop; per-goal override by saying "max mode" / "vanilla" in a brief. 2. **Action mode** — proactive (default) / conservative 3. **Communication** — lead-with-outcome (default) / detailed 4. **Scope discipline** — elaborate-ok, polish welcome (default) / minimal 5. **Memory sharing** — git-committed, team-shared (default) / local-only (gitignored) 6. **Verification** — on (default: a fresh-context verifier must sign off before a goal finishes) / off (checked rubric + saved lessons suffice; faster, less rigorous). Note: rigor=max sets this to a 3-verifier panel automatically. 7. **Harness