Content
75%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 high-quality, information-dense analysis skill body: nearly all content is domain knowledge Claude cannot be assumed to know (thresholds, data sources, weights, signal mappings), organized into a clear five-step workflow with cross-validation steps and a concrete output template. Its only real costs are minor over-explanation of familiar concepts, pseudocode-style signal snippets, and a monolithic single-file layout at ~200 lines.
Suggestions
Trim or drop definitions Claude already knows (contango/backwardation parentheticals, 'negative real rates mean negative holding cost') to tighten token efficiency.
Add brief guidance on what to do when the validation steps disagree (e.g., supply-demand bullish but term structure in deep contango) to close the workflow's feedback-loop gap.
Consider moving the per-commodity indicator and seasonality tables into one-level-deep reference files (e.g., references/oil.md, references/gold.md, references/copper.md) so SKILL.md stays a lean overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with domain-specific value (data sources, thresholds, weights, signal mappings) presented as tables and short snippets, not prose padding. Minor over-explanations of things Claude already knows — e.g., 'Contango (futures > spot, normal market)' and 'negative real rates mean negative holding cost' — keep it just below the lean anchor of 5. | 4 / 5 |
Actionability | Concrete, executable guidance throughout: exact thresholds ('TC < $30/ton = tight', 'spread_ratio > 0.02'), a weighted four-factor gold model, a composite scoring template, and a fill-in output format. Not a 5 because the Python snippets use undefined variables (opec_compliance, china_import_yoy) and are explicitly 'simplified', though the skill does honestly note data must be retrieved externally. | 4 / 5 |
Workflow Clarity | The five-step framework is clearly sequenced (supply-demand → inventory → term structure → seasonality → macro) with built-in validation checkpoints — step 3 'confirms the supply-demand judgment' and step 5 validates against macro, plus a note to cross-check lagged inventory with price/basis behavior. Not a 5 because there is no explicit disagreement-handling guidance for when validation steps contradict each other. | 4 / 5 |
Progressive Disclosure | Well-organized single-file structure with clear section headers and a coherent overview → concepts → framework → output flow; no bundle files exist and no references are buried. Not a 5 because at ~200 lines the per-commodity detail (e.g., seasonality tables, commodity-specific indicator tables) is arguably inline content that could be split into one-level-deep reference files, which the ideal anchor rewards. | 4 / 5 |
Total | 16 / 20 Passed |