Content
72%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is highly actionable and well-structured, with the main gaps being a slightly redundant intro paragraph and the absence of an explicit validation/feedback checkpoint in the batch analysis workflow.
Suggestions
Tighten or cut the intro paragraph that restates the description, letting the Prompt Template lead the body.
Add an explicit validation checkpoint in Step 1 or Step 3 — e.g. confirm the number of records retrieved and flag themes backed by low ARR or a single source before ranking.
Consider a short error-handling note for when CRM loss-reason fields are sparse or unpopulated.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The opening paragraph ('Lost deal notes and churn reasons are scattered across CRM fields… This skill reads all of it, clusters the feedback by theme…') largely restates the description's purpose and could be trimmed; otherwise the body is efficient and free of basic-concept padding. | 2 / 3 |
Actionability | Concrete, executable guidance throughout: named CRM fields to collect, enumerated theme categories, a quantitative ranking rule (occurrences × average ARR affected), an explicit output template, and a filename convention (ChurnAnalysis_YYYY-MM-DD.md). | 3 / 3 |
Workflow Clarity | The five-step sequence is clear and includes a quality threshold ('A theme needs at least 2 occurrences'), but this batch CRM read/analyze operation has no validate→fix→retry checkpoint (e.g. confirming record counts or flagging low-confidence themes), which caps clarity per the batch-operations guideline. | 2 / 3 |
Progressive Disclosure | A self-contained skill with no bundle files and well-organized sections (Prompt Template, Setup, Placeholders, Tips); for a single-purpose skill needing no external references, clear organization is sufficient. | 3 / 3 |
Total | 10 / 12 Passed |