Decide

Choose what to do and preserve why. Do not confuse generating options with making a decision.

Work

1. Frame the decision. Reference the governing intent, specification, grounded state, and relevant findings.
2. Identify real decision drivers. Use only factors that materially distinguish acceptable approaches: correctness, simplicity, reliability, compatibility, performance, security, reversibility, operational burden, cost, time, etc. Avoid decorative scorecards.
3. Generate credible alternatives. Consider more than one serious approach when consequences justify it. Include defer/do-nothing when legitimate; do not manufacture strawmen.
4. Collapse risky assumptions. If the choice hinges on an uncertain claim, investigate or prototype it instead of reasoning indefinitely from assumption.
5. Compare consequences. Examine coupling, failure modes, migration, blast radius, reversibility, maintenance, and what each option makes easier or harder.
6. Prefer the smallest sufficient commitment. When outcomes are comparable, avoid irreversible or speculative complexity.
7. Record the rationale. Capture rejected serious alternatives and the evidence or tradeoff that decided the choice.

Return

• Decision
• Rationale
• Alternatives considered
• Tradeoffs accepted
• Assumptions
• Reconsider when — evidence or conditions that would invalidate the decision.

Failure signals

Watch for choosing the first workable idea, architecture by fashion, fake alternatives, unexplored assumptions that dominate the choice, and reopening settled decisions without new evidence.

Done when

There is enough justified commitment to realize the change without silently re-litigating the same decision, and the conditions that should reopen it are explicit.