Content
93%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.
An exemplary lean skill body: concrete MCP tool invocations with exact parameters, confidence thresholds, a confidence-tiered presentation policy, and a risk-ordered deletion sequence with pre-deletion validation. The only gap is the absence of a post-deletion verification step (re-run build/tests to confirm nothing broke), which keeps workflow clarity at 4.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~38-line body is lean and efficient: parameter lists, thresholds ("confidence >= 0.7", "default is 0.5"), and a deletion-order list with zero padding or concept explanation. Every token is Repowise-specific knowledge Claude doesn't already have — anchor 5. | 5 / 5 |
Actionability | Concrete tool calls with exact parameters and values ("get_dead_code()", "safe_only=true", "kind="unreachable_file"", "directory="src/old/"", "get_risk(targets=["path/to/file"])") are copy-paste ready. Per the rubric's instruction-only scoring note, the absence of code blocks is not penalized when guidance is this actionable — anchor 5. | 5 / 5 |
Workflow Clarity | Clear sequence (retrieve → present → pre-delete checks → ordered deletion) with explicit checkpoints ("Confirm with the user", "Call get_risk(...) to double-check dependents", flag recent git activity), so the destructive-operation cap at 3 does not apply. It falls short of anchor 5 because there is no feedback loop or post-deletion verification (e.g., re-run build/tests after removal). | 4 / 5 |
Progressive Disclosure | Under 50 lines with no need for external references, organized into four well-labeled sections (When to call, How to present, Before deleting, Safe deletion order). Per the rubric guideline, a sub-50-line skill with well-organized sections and no bundle files scores 5. | 5 / 5 |
Total | 19 / 20 Passed |