Content
75%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 strong, code-dense reference that is immediately actionable: nearly every section contains runnable Python or CLI commands, and the organization (data → run → analyze → optimize → validate → debug → export) is easy to navigate. The gaps are polish-level: duplicated boilerplate snippets, one non-executable prose pseudo-command for the MCP data check, and self-contained topics that could be split into reference files.
Suggestions
De-duplicate the repeated DataManager/get_candles and engine.run/print_summary snippets — show them once in Quick Start and reference that pattern elsewhere.
Replace the prose MCP instruction with an actual invocable command or a one-line copy-paste example so the data availability check is executable.
Move Parameter Optimization and Walk-Forward Analysis into a references/ file linked from SKILL.md, and either include scripts/run_backtest.py in the bundle or note it lives in the repo, so the CLI section's referenced path resolves.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and code-first with no explanation of concepts Claude already knows, but there is minor duplication: the DataManager/get_candles and engine.run/print_summary snippets from Quick Start are repeated nearly verbatim in the Data Sources and Single Symbol/Timeframe sections, which could be trimmed — this fits anchor 4 rather than the every-token-earns-its-place bar of anchor 5. | 4 / 5 |
Actionability | Most guidance is executable: the Quick Start is copy-paste ready with imports and a full run, and the metrics, grid-search, walk-forward, and CLI sections give concrete code and commands. Minor gaps keep it below anchor 5: the MCP data check ("Use get_stock_bars for SPY with timeframe 15Min and limit 10") is prose rather than an actual command, and the walk-forward snippet uses a `strategy` variable only defined in an earlier section. | 4 / 5 |
Workflow Clarity | The implied sequence (verify data → run → analyze metrics → optimize → walk-forward validate → export) is logically ordered and includes checkpoints like the data availability pre-check, `print_success_criteria`, and out-of-sample comparison. It falls short of anchor 5 because the sequence is implicit in section order rather than an explicit workflow, and there is no validate→fix→retry loop (the debugging table is issue/fix but not a feedback cycle); it is clearly above anchor 3's merely-listed-steps-with-missing-checkpoints. | 4 / 5 |
Progressive Disclosure | A single-file skill (~183 lines) with clear, well-ordered section headers and no nested references, so navigation is easy. It stays at anchor 4 rather than 5 because the skill exceeds the simple-skill size where inline-only content is ideal — the Parameter Optimization and Walk-Forward sections are natural candidates for a references/ file — and the CLI section references scripts/run_backtest.py, which is not present in the skill bundle. | 4 / 5 |
Total | 16 / 20 Passed |