Content
64%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured, highly actionable skill that provides specific API calls, concrete thresholds, and a detailed output template for B2B account health analysis. Its main weaknesses are the lack of validation/error-handling steps in the workflow (e.g., handling missing data or low-volume accounts) and the monolithic structure that could benefit from splitting the report template and common patterns into separate files. Some best practices like 'show trends, not snapshots' are things Claude can infer.
Suggestions
Add explicit validation checkpoints: what to do when queries return no data, when account history is too short, or when feedback sources are empty — this is critical for a multi-step analytical workflow.
Extract the detailed report template and 'Common Patterns' section into separate referenced files (e.g., REPORT_TEMPLATE.md, PATTERNS.md) to reduce the main skill's token footprint.
Trim 'Best Practices' to only non-obvious guidance — items like 'be specific in recommendations' and 'show trends, not snapshots' are things Claude already understands.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably efficient for its complexity, but the extensive output template is verbose and could be condensed. The 'Common Patterns' and 'Best Practices' sections contain guidance Claude could largely infer. The report template takes up significant token space that could be more compact. | 2 / 3 |
Actionability | Each step specifies exact Amplitude API calls (e.g., `Amplitude:query_dataset`, `Amplitude:get_feedback_sources`, `Amplitude:get_feedback_insights`), concrete metrics to compute (DAU/MAU ratio, WoW changes), specific thresholds for classification (DAU/MAU >40% = Healthy), and a detailed output template. This is highly actionable and specific. | 3 / 3 |
Workflow Clarity | The steps are clearly sequenced (0-5) with logical progression from discovery to reporting. However, there are no explicit validation checkpoints or error recovery steps — e.g., what to do if queries return no data, if the account has insufficient history, or if feedback sources are empty. For a multi-step analytical workflow, these gaps matter. | 2 / 3 |
Progressive Disclosure | The content is well-structured with clear headers and logical sections, but it's a monolithic document with no references to supporting files. The detailed report template and common patterns sections could be split into separate reference files. For a skill of this length (~150+ lines), some content decomposition would improve navigability. | 2 / 3 |
Total | 9 / 12 Passed |