Content
68%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 concrete, code-heavy skill that is largely executable and well-organized, with strong config, sizing, and debugging examples. Its main weaknesses are duplicated MCP command listings and a live-trading workflow whose validation gates (success criteria, order verification) are implied rather than explicit.
Suggestions
Add explicit validation checkpoints to the Live Trading Workflow: define the backtest success criteria (e.g., win rate / max drawdown thresholds) and a paper-trading gate before 'Scale up', plus a post-order verification step (confirm fill via get_all_orders / get_all_positions) given the batch and liquidation commands (close_all_positions, cancel_all_orders).
Deduplicate the MCP commands: the 'Paper Trading Commands' block under Live Trading Workflow repeats get_account_info, get_all_positions, get_open_position, place_stock_order, and close_position from the 'Paper Trading with MCP' section — keep one listing and cross-reference it.
Tighten the Entry Logic Deep Dive so snippets are runnable: bind confirmation_ratio (e.g., from config) and show the analyzer/detector construction, mirroring the self-contained debug_entry example.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense code and config with no concept padding, but MCP commands are listed twice ("Use get_account_info to view current balance and buying power" / "# Check account / Use get_account_info" and repeated get_all_positions/close_position lines), a minor trim opportunity — efficient with minor redundancy rather than noticeably verbose (rule out 3). | 4 / 5 |
Actionability | Mostly executable: copy-paste config dict, worked sizing example ("shares = 1000 / 1.50 # 666 shares"), complete debug_entry function, and a Common Issues table; however the Entry Logic Deep Dive snippets use unbound names ("confirmation_ratio", fragmentary "self.trend_detector" calls) — minor gaps keep it below fully copy-paste ready (rule out 5). | 4 / 5 |
Workflow Clarity | The Live Trading Workflow lists a sequence with one implicit gate ("Backtest first - Ensure strategy passes success criteria") but the success criteria are never defined, and consequential/batch operations ("Use close_all_positions to liquidate everything", "cancel_all_orders", live order placement) lack explicit validation or verification checkpoints — per the rubric cap this cannot exceed 3 (rule out 4: checkpoints are missing, not just minor-gapped). | 3 / 5 |
Progressive Disclosure | No bundle files exist, and the single file is well-organized with clear section headers; the ~40 lines of MCP command catalog (duplicated across two sections) is the only content that arguably belongs in a separate reference file — good structure with a minor organization gap (rule out 3: nothing is buried or badly placed; rule out 5: the command catalog could be split out and deduplicated). | 4 / 5 |
Total | 15 / 20 Passed |