Content
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides a well-structured framework for defining North Star Metrics with a clear multi-step process and good evaluation criteria. However, it is significantly over-verbose, spending hundreds of tokens explaining product strategy concepts that Claude already knows. It would benefit from aggressive trimming of the Domain Context section, adding a concrete worked example, and defining explicit feedback loops when validation fails.
Suggestions
Cut the 'Domain Context' section by 70%+ — remove explanations of what NSMs are/aren't and basic product concepts Claude already knows. Keep only the Three Games classification and the Metric System layers as brief reference.
Add a concrete worked example showing a sample product analyzed through all 6 steps with specific scores, so Claude has a clear template for output quality.
Define an explicit scoring scale for the six quality criteria in Step 3 (e.g., 1-5) and show what a scored candidate looks like.
Add an explicit feedback loop in the workflow: if Step 6 validation reveals structural weaknesses, specify returning to Step 2 or Step 4 to iterate on candidates.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose, explaining concepts Claude already understands well (what a North Star Metric is, what it isn't, what OKRs are, what vanity metrics are). The 'Domain Context' section alone is several hundred words of background that Claude doesn't need. The 'What a North Star Metric is NOT' section explains basic product strategy concepts. Much of this could be cut to a fraction of its length. | 1 / 3 |
Actionability | The prompt section provides a clear 6-step workflow with specific evaluation criteria and output format, which is reasonably actionable. However, there's no concrete example of a completed analysis (e.g., a sample product with its NSM, input metrics, and scoring), and the scoring criteria lack a defined scale (1-5? 1-10?). The guidance is structured but not fully executable. | 2 / 3 |
Workflow Clarity | The 6-step process is clearly sequenced and Step 6 provides a validation checkpoint. However, there's no feedback loop for what to do if validation fails (e.g., if the NSM doesn't predict revenue, go back to Step 2). For a framework that explicitly warns about getting it wrong, the lack of explicit error recovery or iteration guidance is a gap. | 2 / 3 |
Progressive Disclosure | The content is organized with clear sections and headers, and references external reading at the end. However, the Domain Context section is a large inline block that could be separated or significantly condensed. There are no bundle files to offload detailed content to, and the skill mentions pairing with 'build-metric-tree' but doesn't link to it. The monolithic structure could benefit from splitting. | 2 / 3 |
Total | 7 / 12 Passed |