Content
62%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a highly actionable and well-structured skill with an excellent 7-phase workflow, concrete commands, and a thorough output template. However, it suffers significantly from verbosity — it explains IaC syntax Claude already understands, includes inline content that should be in referenced files, and over-specifies discovery heuristics. The progressive disclosure structure exists (7 reference files) but isn't leveraged enough, leaving the main file bloated.
Suggestions
Move the estimation formulas table, pricing lookup examples, and the full output report template to referenced files (e.g., service-mapping.md and cost-projection-output-examples.md are already referenced — use them instead of duplicating content inline).
Remove explanations of IaC syntax (what Terraform resource blocks look like, what CDK construct instantiations look like, what CloudFormation Resources sections look like) — Claude already knows these formats.
Consolidate the 5-step auto-discovery process into a concise checklist rather than showing every grep/find command; move the detailed commands to a reference file.
Remove or drastically shorten the 'File Pattern → Framework' table and IaC parsing examples — a single sentence like 'Detect framework from file extensions and imports, then extract resource types, names, and cost-relevant configuration' suffices.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~400+ lines. It over-explains discovery heuristics (5 sub-steps for auto-discovery), includes lengthy estimation tables Claude could look up, explains what CDK/Terraform/CloudFormation resource blocks look like (Claude already knows), and repeats security/best-practice guidance that could be in referenced files. Significant token waste. | 1 / 3 |
Actionability | The skill provides highly concrete, executable guidance throughout: specific bash commands for file discovery, exact API call parameters (group_by, filters, granularity), web search query templates, pricing formulas with specific numbers, and a complete output template with example values. Copy-paste ready at every step. | 3 / 3 |
Workflow Clarity | The 7-phase workflow is clearly sequenced with explicit goals for each phase. Phase 3 includes validation steps (confirming dimension values exist in CloudZero), Phase 4 specifies how to calculate run rates, Phase 5 has a prioritized strategy cascade (web lookup → similar resources → fallback formulas), and Phase 6 has explicit aggregation logic. Confidence levels serve as validation checkpoints for estimates. | 3 / 3 |
Progressive Disclosure | The skill references 7 external files in 'See Also' and uses ${CLAUDE_PLUGIN_ROOT}/references/ paths throughout, which is good structure. However, the main SKILL.md itself is monolithic — the estimation formulas table, the full output template, the 5-step auto-discovery process, and the IaC parsing details could all be in referenced files. Too much detail is inline rather than delegated to the referenced documents. | 2 / 3 |
Total | 9 / 12 Passed |