Content
72%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 well-organized, actionable sentiment-analysis reference with concrete thresholds and a usable scoring/output template; its main weakness is the absence of explicit validation checkpoints in the analysis-to-action workflow.
Suggestions
Add an explicit workflow with checkpoints: compute each dimension → cross-validate (e.g. PCR vs VIX, 融资 vs 北向) → flag conflicting signals → only then map to a position.
De-duplicate the 0-100 sentiment-scale mapping (currently restated in 恐贪, 综合 scoring, and output sections) into one reference table.
Tighten pedagogical asides like the '巴菲特指标' anecdote and '别人恐惧时贪婪' to keep the body lean.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense, table-driven content assumes Claude's domain knowledge and mostly earns its tokens, though the ~270-line body repeats the 0-100 sentiment scale mapping across sections and includes a few pedagogical asides. | 4 / 5 |
Actionability | Concrete thresholds, specific signal rules, a ready scoring formula ('综合情绪 = 0.25×恐贪 + 0.20×PCR + ...'), and a copy-paste output template are actionable; minor gaps are no executable data-fetch commands behind the named data sources. | 4 / 5 |
Workflow Clarity | The scoring-to-position mapping is a clear sequence and the notes list caveats, but there are no explicit validation checkpoints or a validate→fix→retry loop, leaving checkpoints implicit. | 3 / 5 |
Progressive Disclosure | A single self-contained SKILL.md with well-organized section headers covering each sentiment dimension and an output format; with no external references needed this structure is easy to navigate. | 5 / 5 |
Total | 16 / 20 Passed |