Content
68%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is well-structured and highly actionable, with a complete output template and explicit routing. Its main weakness is workflow clarity: the retrospective process is implied by the record template rather than stated as steps, and no validation or fallback checkpoints exist for missing outcome data.
Suggestions
Add an explicit numbered procedure before the output template: locate the decision record and execution plan, gather actual outcomes, score against pre-committed criteria, audit assumptions and dissent, then write the record — instead of relying on the template's section order to imply the sequence.
Add validation/branching checkpoints, e.g., what to do when actual outcome metrics are unavailable (mark criterion UNMEASURED and note the gap) or when the pre-committed criteria cannot be found in the decision record.
Define the rule for choosing the overall WIN/PARTIAL/LOSS/MIXED status (e.g., based on which success criteria were met vs. which kill criteria triggered) so the status is scored rather than guessed.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is largely lean: the output template is the deliverable itself and sections are short and functional. Minor over-explanation remains — the "Why Pre-Committed Criteria Matter" and "Why Revisit Dissent" rationale paragraphs and the line "This is the rigor that compounds over time" could be trimmed — matching anchor 4 rather than the every-token-earns-its-place anchor 5, and well above the some-unnecessary-explanation anchor 3. | 4 / 5 |
Actionability | Provides a copy-paste-ready output template with an exact save path (`~/.claude/postmortems/YYYY-MM-DD-<slug>.md`), concrete scoring tables, dissent/assumption audit formats, and explicit routing commands. Not 5 because minor gaps remain: no decision rule for choosing WIN/PARTIAL/LOSS/MIXED overall status, and no pointer to where the input decision record is stored. | 4 / 5 |
Workflow Clarity | The sequence is present but implicit — conveyed by the template's section order (Outcome Scoring → Dissent → Assumption Audit → Process Lessons → Status) rather than explicit procedural steps, and validation checkpoints are missing (e.g., what to do when actual outcome metrics are unavailable, or how to verify the decision record was found). This matches anchor 3 ("sequence present but checkpoints missing or implicit"); not 4 because execution steps are never stated as a process, though the artifact structure keeps it above anchor 2. | 3 / 5 |
Progressive Disclosure | A self-contained single-file skill with no bundle files, well-organized sections, and clearly signaled cross-links (Routing, Related). Not 5 because the ~55-line output template makes up half the file and could live in a reference file, and at ~110 lines the skill exceeds the under-50-line threshold where well-organized sections alone merit a 5; not 3 because the structure is good and no content is clearly misplaced. | 4 / 5 |
Total | 15 / 20 Passed |