Content
86%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.
The body is token-efficient and well-structured, with concrete formulas and clear signal logic. Its main gap is the absence of a fully runnable example and output-validation guidance.
Suggestions
Add a short, complete copy-paste example that constructs the indicator on a sample OHLCV DataFrame and emits the composite signal column.
Note warmup/edge-case handling (e.g. required history length for EMA/ADX/BB, NaN behavior in the first rows) so outputs can be validated.
Clarify how to verify the signal is sensible (e.g. sanity checks against known trending vs. ranging segments) before acting on it.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean tables and bullets with no padding; it surfaces only non-obvious detail (e.g. 'RSI and ADX use Wilder EWM (ewm(alpha=1/period)), not a rolling mean') and assumes Claude's competence. | 5 / 5 |
Actionability | Gives concrete, executable specifics — exact pandas formulas like 'OBV = (volume * sign(close.diff())).cumsum()' and the full ADX chain — but lacks a complete copy-paste runnable signal-generation example. | 4 / 5 |
Workflow Clarity | The signal logic is clearly sequenced via the three-dimensional voting conditions (Long/Short/Stand aside), but there are no validation or edge-case checkpoints (e.g. NaN handling, insufficient warmup length). | 4 / 5 |
Progressive Disclosure | A compact (~50 line), single-purpose skill with no external references needed; well-organized into Purpose, Signal Logic, Implementation Details, Parameters, Dependencies, and Signal Convention sections. | 5 / 5 |
Total | 18 / 20 Passed |