What If

Use this skill for counterfactual planning, not implementation-first work.

Workflow

1. Frame the thought experiment.
   • State the current world in 3-6 bullets.
   • State the changed constraint(s) explicitly.
   • State the invariants: what must still be true.
   • If any of those are fuzzy, ask a few focused questions first.
2. Sketch 2-4 plausible alternative worlds.
   • Include at least one conservative option and one more radical option.
   • Do not invent fantasy paths that ignore the user's real tooling, migration cost, or constraints.
   • For each world, describe the shape, the implementation outline, the upside, and the downside.
3. Stress-test each world adversarially.
   • Load references/review-lenses.md.
   • Look for breakpoints, hidden coupling, migration traps, security/trust issues, operator burden, and reversibility.
   • If subagents or parallel review tools are available, use them for an adversarial pass; otherwise do the adversarial review yourself.
4. Compare the worlds and make a call.
   • Say which world is best, which is safest, and which is most interesting-but-dangerous.
   • It is valid to recommend staying with the current setup if the alternatives are not compelling.
5. Capture the findings.
   • Use the output shape in references/review-lenses.md unless the user wants a different format.
   • Keep the final result decision-oriented, not stream-of-consciousness brainstorming.

Working Rules

• Prefer concrete consequences over abstract opinions.
• Separate facts, assumptions, and speculation.
• Name second-order effects, not just first-order benefits.
• Include migration steps when proposing a new world.
• Include rollback or coexistence paths when relevant.
• Do not force a change; "keep the current tradeoff" is a valid outcome.