Content
71%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-structured skill body with excellent progressive disclosure — four real one-level-deep references, each signaled inline and indexed with When/Contains entries — and clearly sequenced workflows with explicit use preconditions. The main weaknesses are inline duplication of standard finance knowledge (RSI thresholds, chart patterns, generic formatting advice) that belongs in or stays redundant to the reference files, and instructions like 'estimate fair value range' that defer method without showing a worked example.
Suggestions
Remove textbook finance knowledge from the body — RSI thresholds, MACD crossover meanings, standard chart pattern names, and the generic 'Output Guidelines'/'Tone' formatting advice — since references/technical-analysis.md and references/financial-metrics.md already cover these; keep only the workflow steps that direct what to fetch and which reference to read.
Add an explicit validation checkpoint in the data-gathering step, e.g., 'If sources disagree on a metric by >10%, note the discrepancy and prefer the most recent SEC filing' — currently 'verify data recency' is the only data-quality check before analysis begins.
Include one compact worked example in the body (a small filled-in metrics table or a 3-line fair-value calculation) so instructions like 'Estimate fair value range' and 'Create side-by-side comparison' have a concrete output shape to follow.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient as structured workflow steps, but includes explanations Claude already knows: 'RSI: Overbought (>70) or oversold (<30)', 'Reversal patterns (head and shoulders, double top/bottom)', and generic formatting/tone advice ('Use tables for financial data', 'Objective and balanced', 'Avoid hyperbole') that duplicates content in references/technical-analysis.md and adds no skill-specific value. This matches anchor 3 ('mostly efficient but includes some unnecessary explanation or could be tightened') rather than anchor 2, since there is no padded prose and the structure is sound — but the knowledge-duplication is more than 'minor', keeping it below anchor 4. | 3 / 5 |
Actionability | As an instruction-only skill, it gives concrete guidance: exact search strings ('AAPL financial metrics 2024'), enumerated data to fetch, named quality sources, per-workflow reference reads, and defined output formats. It falls short of anchor 5 because some instructions remain high-level — 'Estimate fair value range' and 'Develop bull and bear cases' defer method entirely to references, and no filled-in example output or table is shown — matching anchor 4 ('mostly executable guidance; concrete code or commands with minor gaps'). | 4 / 5 |
Workflow Clarity | Four workflows are clearly sequenced with 'When to Use' preconditions and stepwise numbering, and workflows compose (the comprehensive report reuses the fundamental and technical workflows). It matches anchor 4 ('clear sequence with most checkpoints present; minor validation gaps'): the only validation checkpoint is 'Always verify data recency (prefer data from last quarter)', with no explicit cross-check of conflicting data between sources or verify-completeness step before writing the report. It is above anchor 3, which requires checkpoints to be missing or implicit — the recency check is explicit — and the operations are read-only, so no destructive/batch validation cap applies. | 4 / 5 |
Progressive Disclosure | The body is an overview plus workflows that point to four real, one-level-deep reference files (all verified present in references/: technical-analysis.md, fundamental-analysis.md, financial-metrics.md, report-template.md), each signaled inline at the point of need ('Read references/technical-analysis.md for indicator definitions') and summarized in a 'Reference Files' section with When/Contains entries, with no nested references. This matches anchor 5 ('Clear overview with well-signaled one-level-deep references; content appropriately split; easy navigation'). | 5 / 5 |
Total | 16 / 20 Passed |