菜单
Use when an agent needs a standard cross-sectional long-only backtest protocol with T+1 open execution, top-bucket equal weighting, fees, limit-up or limit-down exclusions, benchmark comparison, NAV curves, drawdown, IC, and diagnostic charts.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/XinAloha/skills --skill backtest命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Brinson-Fachler / BHB sector attribution with interaction, concentration (HHI), top contributors, and optional Carino multi-period linking. Use to explain active return as allocation vs selection.
Run and explain cross-sectional trading-factor backtests on schema-compatible factor signals and market data. Use when Codex needs to validate backtest inputs; calculate returns, NAV, drawdown, holdings, turnover, benchmark or hedged performance; generate a PDF diagnostic report; or explicitly verify the checked-in test fixture. Also use for Chinese requests such as 因子回测、收益验证、净值、回撤、持仓、换手、对冲收益或回测报告.
Decompose a portfolio's active return versus its benchmark into sector allocation, stock selection, and interaction (Brinson-Fachler + Carino multi-period linking) plus factor contributions (cross-sectional regression), answering where the excess return came from. Evidence-first with built-in identity assertions. Use when the user asks 组合归因, 业绩归因, 超额收益来源, 这策略赚的是配置还是选股, or 回测跑完想知道钱赚在哪, on Claude Code, Codex, Cursor, Hermes, or OpenClaw.
正在显示 SKILL.md
| name | backtest |
| description | Use when an agent needs a standard cross-sectional long-only backtest protocol with T+1 open execution, top-bucket equal weighting, fees, limit-up or limit-down exclusions, benchmark comparison, NAV curves, drawdown, IC, and diagnostic charts. |
| quantSkills | {"project_type":"skill","category":"tooling","tags":["backtest","cross-section","long-only","diagnostics","quant-research"],"platforms":["claude-code","codex","openclaw","cursor"],"status":"stable","validation_level":"listed","maintainer_type":"community","summary_zh":"不是回测框架，而是截面多头回测的标准协议：T+1 开盘成交、Top 等权、双边 15bp、涨跌停剔除、四联诊断图、5 项健康度自检。","summary_en":"Standard cross-sectional long-only backtest protocol with T+1 execution, fees, limit filters, NAV curves, IC, drawdown, and diagnostic charts.","license":"GPL-3.0"} |
{
"version": 1,
"task": {
"placeholder": "补充信号文件、回测区间、基准或特殊假设（可选）"
},
"fields": [
{
"key": "factor",
"label": "内置因子",
"type": "select",
"default": "momentum_20",
"help": "填写自定义表达式时以表达式为准",
"options": [
{ "value": "momentum_20", "label": "20日动量" },
{ "value": "reversal_5", "label": "5日反转" },
{ "value": "lowvol_20", "label": "20日低波动" },
{ "value": "alpha101_101", "label": "Alpha101 #101" },
{ "value": "alpha101_12", "label": "Alpha101 #12" },
{ "value": "corr_open_vol", "label": "量价背离" }
]
},
{
"key": "expr",
"label": "自定义因子表达式",
"type": "textarea",
"placeholder": "例如：-1 * correlation(rank(open), rank(volume), 10)"
},
{
"key": "universe",
"label": "股票池",
"type": "select",
"default": "000300.SH",
"options": [
{ "value": "000300.SH", "label": "沪深300" },
{ "value": "000905.SH", "label": "中证500" },
{ "value": "399006.SZ", "label": "创业板指" },
{ "value": "000852.SH", "label": "中证1000" }
]
},
{
"key": "horizon",
"label": "持有周期",
"type": "select",
"default": "5",
"options": [
{ "value": "1", "label": "1日" },
{ "value": "5", "label": "5日" },
{ "value": "10", "label": "10日" }
]
}
],
"prompt_template": "{{#task}}任务与材料：\n{{task}}\n\n{{/task}}{{#attachments}}用户上传的材料（已放入工作区）：\n{{attachments}}\n\n{{/attachments}}对 {{universe}} 中的因子 {{factor}}{{#expr}}（自定义表达式优先：{{expr}}）{{/expr}} 按 {{horizon}} 日持有周期执行标准截面多头回测，严格采用 T+1 开盘成交、Top 10% 等权、滚动持仓、双边费用、涨跌停与停牌过滤，比较基准并给出净值、回撤、换手、IC、分层收益和健康度诊断，输出中文报告。"
}
把信号
[date × symbol]模拟交易出来，得到净值 / 回撤 / 换手 / 分组收益。从"统计相关"走向"可执行收益"的关键一步。
references/assumptions.md）trade_status==1 或 close >= limit_up*0.99references/diagnostic-charts.md）references/health-check.md）1. 校验信号契约（截面规模 / std / 涨跌停字段是否齐）
2. 算 T+1 开盘成交矩阵（mask 涨跌停 / 停牌）
3. 滚动持仓回测（队列 sleeves，每日 1/H 换仓）
4. 算指标（年化 / Sharpe / MDD / 换手）
5. 出四联图 + benchmark 对齐
| 本 skill 概念 | 你的项目对应 |
|---|---|
panel 必须字段 | open / close / high / low / volume / trade_status / limit_up / limit_down |
signal | [date × symbol] 浮点 DataFrame，已截面 z-score |
horizon | 必须等于信号声明的 H |
项目内的 backtest() | 永远优先调用，不要自己另写一份"通用版"（破坏评估口径） |
| 何时读 | 文件 |
|---|---|
| 不确定回测假设是不是合理 | references/assumptions.md |
| 想看核心算法骨架 | references/algorithm.md |
| 不会画四联图 | references/diagnostic-charts.md |
| 跑完想做健康度自检 | references/health-check.md |
| 怀疑结果不对劲 | references/anti-patterns.md |
agents/openai.yamlagents/cursor-rule.mdcagents/portable-loader.md按 QUANTSKILLS 社区规则 §8 声明。