Content
67%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 dense, well-structured reference with genuinely concrete trade examples and risk-control checkpoints. Its weaknesses are the re-explanation of basic options concepts Claude already knows and the lack of executable calibration/analysis code despite declared Python dependencies.
Suggestions
Cut or compress explanations Claude already knows, e.g. the 'option price = intrinsic value + time value' framing and the first-order Greeks meaning table, keeping only the management-approach column.
Provide executable code for at least one workflow (e.g. SABR calibration or a Greeks-monitoring script) matching the declared pandas/numpy/scipy dependency, rather than prose-only description.
Move deep-dive material (SABR calibration procedure, market-making operations, per-strategy details) into one-level-deep reference files and keep SKILL.md as a lean overview with clearly signaled links.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Valuable domain-specific content (SABR parameter ranges, China skew direction, Zakamouline hedge trigger) is diluted by explanations Claude already knows, e.g. 'option price = intrinsic value + time value' and the first-order Greeks definitions table, so it could be tightened. | 3 / 5 |
Actionability | Concrete numbered trade examples (50ETF premiums, IVs, breakeven ranges), formulas like 'Breakeven volatility = IV + Theta/Gamma cost', and explicit numeric risk thresholds make guidance mostly executable; only minor gaps remain, e.g. SABR calibration is described but no code is given despite the pandas/numpy/scipy dependency. | 4 / 5 |
Workflow Clarity | Each strategy follows a clear entry-conditions → trade → risk-control sequence with explicit checkpoints ('stop loss if loss exceeds 50% of net debit', 'term-structure inversion → close position'), though no end-to-end surface-analysis-to-recommendation sequence or error-recovery loop is provided. | 4 / 5 |
Progressive Disclosure | The single file is well organized with clear section headers, tables, and code blocks; however, at ~185 lines, deep-dive material (SABR calibration details, market-making operations, per-strategy playbooks) could be split into one-level-deep reference files. | 4 / 5 |
Total | 15 / 20 Passed |