Content
71%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.
A well-structured, token-efficient overview with a clear workflow and explicit completion gates, appropriately sized for a skill with no bundle files. Its weakness is actionability: it names the right analytical operations but delegates the actual formulas and specifics rather than providing them.
Suggestions
Add the key formulas explicitly (e.g., NRR, gross churn, CAC payback, net burn / runway months) or point to a reference file containing them, so 'transparent formulas' is specified rather than delegated.
Add a feedback loop for reconciliation failures, e.g., 'If totals do not reconcile, re-check data grain and churn definition before proceeding.'
Remove the '## Purpose' section that duplicates the frontmatter description verbatim to recover tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean with no concept-teaching padding, but the '## Purpose' section repeats the frontmatter description verbatim ('Analyze SaaS growth, retention, efficiency, unit economics, runway, and reporting quality from supplied data.') and the copyright line adds nothing for Claude — minor instances that could be trimmed. Better than 3 because there is no unnecessary explanation; short of 5 because of the verbatim duplication. | 4 / 5 |
Actionability | No concrete formulas, commands, or example calculations are provided anywhere: 'Calculate retention, growth, efficiency, and runway with transparent formulas' delegates the formulas rather than specifying any (no NRR, CAC payback, burn multiple, or Rule of 40 definitions), and 'Prioritize actions by likely economic impact and confidence' is abstract. Better than 2 because the workflow steps and completion gates are specific and sequenced rather than high-level hints. | 3 / 5 |
Workflow Clarity | A clear 5-step sequence (clarify, reconcile, calculate, segment, prioritize) with explicit validation checkpoints in 'Completion gates' ('Reconcile totals back to source data', 'State every formula and denominator'). Not 5 because there is no validate-fix-retry feedback loop (e.g., what to do when totals do not reconcile); better than 3 because checkpoints are explicit rather than implicit. | 4 / 5 |
Progressive Disclosure | The body is under 50 lines, needs no external references, and is organized into well-labeled sections (Purpose, Tool routing, Workflow, Completion gates, Boundaries, Delivery), which the rubric guideline says earns a 5 for simple skills with well-organized sections. | 5 / 5 |
Total | 16 / 20 Passed |