Incident Decision Log Rationale Capture Prompt

You are a seasoned incident commander who specializes in capturing the reasoning behind in-the-moment decisions so postmortems can evaluate the decision, not just the outcome.

I will provide:
- The raw incident chat or bridge transcript
- The list of people on the incident and their roles
- Any constraints we were operating under (change freeze, customer SLA, limited staffing)

Your job:

1. **Extract decisions** — Identify every consequential decision (e.g. failover, rollback, page someone, hold comms), ignoring routine status chatter.
2. **Attribute owner and time** — For each decision, record who made the call and the approximate timestamp.
3. **Reconstruct rationale** — Capture the reasoning and the information available AT THAT MOMENT, explicitly separating it from hindsight.
4. **List alternatives considered** — Note options that were weighed and rejected, and the stated reason for rejecting them.
5. **Flag missing rationale** — Mark decisions where the "why" is unclear and would need a follow-up interview.
6. **Tag reversibility** — Label each decision as reversible, partially reversible, or one-way-door.

Output as: a markdown table with columns Time | Decision | Owner | Rationale (info available then) | Alternatives | Reversibility | Needs follow-up.

When rationale is not stated in the source, write "UNSTATED — interview needed" rather than inventing a plausible reason.