Content
78%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 an impressively lean, domain-specific framework with concrete tables, ratios, and executable discovery commands organized in a clear step sequence. Its main weaknesses are the missing validation checkpoints and its reliance on a docs/output-kit.md format spec (and /atlas-report) that are not bundled with the skill.
Suggestions
Bundle docs/output-kit.md (or inline the output-format spec) so the required 40-line CLI format and severity indicators are actually available; the same applies to clarifying what /atlas-report is and when it exists.
Add explicit validation checkpoints before Delivery, e.g., verify tier % of accounts and % of ARR each sum to ~100%, and check that the per-tier CSM counts in the resource map are consistent with the stated CSM:account ratios.
Provide one worked example with real numbers (e.g., a filled threshold set for a $10M-ARR company) so the $[X]/$[Y]/$[Z] and [N] placeholders have a concrete reference point.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense, prescriptive domain content — health-score weight tables, CSM ratios, cadences, and promotion rules — with essentially no explanation of concepts Claude already knows and only trivial glosses (the NRR parenthetical). It assumes Claude's competence and every section carries information Claude cannot supply itself, matching the lean-and-efficient anchor. | 5 / 5 |
Actionability | Provides executable find/grep commands in Step 0 and concrete numbers (weights, score ranges, CSM ratios, cadences) throughout. Not 5 because the tier tables and resource map are placeholder templates ($[X], $[Y], [N]) rather than filled examples, and the required output format depends on a docs/output-kit.md file that is not in the bundle; well above 3 since real commands and specific parameters are given. | 4 / 5 |
Workflow Clarity | Steps 0-6 are clearly sequenced with a distinct purpose each, and Step 0's 'Ask for missing inputs' acts as an input checkpoint. Not 5 because there are no validation steps on the outputs (e.g., verifying tier percentages sum to ~100%, or CSM headcount matches the stated ratios); not 3 because the sequence is complete and checkpoints are mostly present, and this is a document-generation task so the destructive/batch cap does not apply. | 4 / 5 |
Progressive Disclosure | Sections are well-organized, but the body references 'docs/output-kit.md' and '/atlas-report', neither of which exists in the skill's bundle (no references/, scripts/, or assets/ directories at all), leaving dangling navigation. At ~140 lines with all content inline, the under-50-line no-references exception does not apply; this sits at 'some structure, references present but not resolvable' rather than the well-placed-content of 4. | 3 / 5 |
Total | 16 / 20 Passed |