Content
88%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.
A comprehensive, well-structured analytical skill body with concrete actionable guidance, a clear sequenced workflow, and useful checklists. The main weakness is that everything lives inline in one long file with no progressive disclosure to reference materials.
Suggestions
Move the detailed 'Specific Check Guidance' and 'Core Checks' catalogs into a reference file (e.g., references/checks.md) and link to it from SKILL.md to reduce body length and improve progressive disclosure.
Tighten the overlap between the 'Run core quality checks' workflow step and the 'Core Checks'/'Specific Check Guidance' standards sections to remove redundant framing.
Add a brief worked SQL or Python profiling snippet (row count, null rate, duplicate rate) in the 'Build a compact profile' step to make the most common first action copy-paste executable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense, specialized guidance that assumes Claude's competence and avoids explaining basic concepts, with only minor redundancy between the 'Run core quality checks' and 'Specific Check Guidance' sections that could be tightened. | 4 / 5 |
Actionability | Highly concrete, actionable instruction guidance — exact checks, sentinel values ('', 'unknown', 'n/a', 0, -1), cross-field contradictions (is_cancelled = false with non-null cancelled_at), shape-specific check lists, and explicit tool routing ($jupyter-notebooks, $build-report, structured_data) — covering the common cases. | 5 / 5 |
Workflow Clarity | A clearly sequenced 8-step workflow with explicit checkpoints ('Confirm grain before interpreting anomalies', 'infer cautiously and label assumptions'), supporting checklists (Core Checks, shape-specific checks), and a severity rubric that functions as validation for findings. | 5 / 5 |
Progressive Disclosure | Well-organized into clear sections (Workflow, Standards, Defaults) with no nested references, but the body is a single ~160-line monolith with no offloaded reference files, so detailed check catalogs are inlined rather than split out. | 4 / 5 |
Total | 18 / 20 Passed |