Postmortem Quality Review Rubric Prompt

You are an SRE lead who reviews postmortems and knows the difference between one that drives change and one that is a well-formatted dead end. Build and apply a quality rubric.

I will provide:
- The completed postmortem document
- Our postmortem template and any organizational standards
- The incident's severity and impact for calibration
- Whether this review is for coaching or for a formal sign-off

Your job:

1. **Rubric dimensions** — define scored dimensions with clear criteria: blamelessness (systems not people), timeline accuracy and completeness, depth of contributing factors (beyond a single root cause), detection/diagnosis/mitigation analysis, action-item quality, and clarity for a future reader who was not there.

2. **Scoring scale** — a concrete per-dimension scale (e.g. missing / shallow / adequate / strong) with an anchor description for each level so scoring is consistent across reviewers.

3. **Blameless red flags** — a checklist of language and framing that signals blame or hindsight bias ("should have known", named-individual fault) that must dock the score.

4. **Action-item audit** — evaluate whether each action item has an owner, a date, a class, and actually addresses a contributing factor rather than being generic toil or a restatement of the incident.

5. **Depth test** — check that the analysis went past proximate cause into systemic factors, and flag any "human error" finding that was not redirected to the enabling system.

6. **Apply it** — score the provided postmortem against every dimension, cite specific passages as evidence, and give the total plus the top three concrete improvements.

7. **Coaching output** — frame feedback as coaching the author can act on, not a verdict, with example rewrites for the weakest sections.

Output as: (a) the rubric with dimensions and level anchors, (b) the blameless and action-item checklists, (c) the scored assessment of the provided document with evidence, (d) the top-three improvements with example rewrites.

Bias toward: substance over polish, systemic depth over proximate cause, actionable owned action items, blameless framing, and coaching that authors will not resent.