Content
57%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 a dense, actionable reference with concrete thresholds, code, and China A-share specifics, but it is a monolithic single file with no progressive disclosure and only an implied workflow rather than an explicit sequenced procedure.
Suggestions
Add an explicit step-by-step analysis workflow (e.g. 1. score sentiment 2. detect biases 3. select strategy 4. apply debiasing checklist) with the Output Format as the validation/template checkpoint.
Move the detailed cognitive-bias tables and sentiment-indicator catalogs into a references/ file linked from the overview to improve progressive disclosure.
Trim definitions of well-known biases (loss aversion, anchoring) and keep only the quant-detection and debiasing columns.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly information-dense with China A-share-specific thresholds and empirical figures that earn their place, but it explains basic biases (loss aversion, anchoring) Claude already knows and could be tightened. | 3 / 5 |
Actionability | Provides concrete executable guidance — numeric signal thresholds (SUE>2σ, RSI<10), a Python sentiment-score skeleton, and a CGO formula — with minor gaps from undefined helpers (normalize, weighted_sum). | 4 / 5 |
Workflow Clarity | The Output Format and numbered Analysis Framework imply a sequence (diagnose sentiment, detect biases, recommend strategy, debias), but there is no explicit step-by-step procedure with validation checkpoints. | 3 / 5 |
Progressive Disclosure | Well-organized into clear sections, but as a >50-line monolithic file with no bundle references, detailed reference material (bias tables, indicator catalogs) that could live in separate files is inlined. | 3 / 5 |
Total | 13 / 20 Passed |