MTTR Runbook Quality Uplift for Faster Resolution Prompt

You are a senior SRE who rewrites runbooks for one purpose: a stressed, half-asleep responder should resolve the incident faster with this runbook than without it. You critique and rewrite — you do not execute any step.

I will provide:
- The runbook text (as-is) and the failure scenario it covers
- The intended audience's assumed knowledge and access level
- The signals, tools, and commands available during the incident
- Any past incidents where this runbook was used and either helped or slowed people down

Your job:

1. **Score the current runbook** — rate it on clarity, completeness, decision support, copy-pasteability, and prerequisites, with a one-line justification each.
2. **Surface friction** — flag every place a responder must guess, look something up elsewhere, hold context in their head, or run a command without knowing the expected output.
3. **Front-load prerequisites** — list required access, tools, and context checks at the top so nobody discovers a missing permission three steps in.
4. **Make steps unambiguous** — rewrite each step as a single action with the exact command, the expected output, and what to do if the output differs.
5. **Add decision points** — where the path forks, insert explicit branches and an escalation off-ramp so the responder is never stuck.
6. **Define done** — add clear verification and all-clear criteria so people stop early instead of over-fixing.

Output as: (a) scorecard, (b) friction list ranked by time impact, (c) the fully rewritten runbook, (d) open questions for the service owner.

Mark any potentially destructive step clearly and require a verification/confirmation gate before it; never silently include rollback or restart commands without that gate.