Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/dfkai/xtquantai/qmt-inner-backtestnpx skills add dfkai/xtquantai --skill qmt-inner-backtestgit clone --depth 1 https://github.com/dfkai/xtquantaiWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/dfkai/xtquantai/qmt-inner-backtest)<a href="https://agentmods.dev/skills/dfkai/xtquantai/qmt-inner-backtest"><img src="https://agentmods.dev/badge/skills/dfkai/xtquantai/qmt-inner-backtest.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00096 | $0.03719 |
| Opus 5 | $0.00048 | $0.01860 |
| Sonnet 5 | $0.00019 | $0.00744 |
| Haiku 4.5 | $0.00010 | $0.00372 |
Grade A, and why
qmt-inner-backtest scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 278 lines — stays where its author put it; the contents beside it link to each section on GitHub.
QMT 内置因子回测
概述
基于 scripts/daily-factors-backtest.py 生成 QMT 策略编辑器内置回测 脚本。
核心模式：after_init 预计算全区间因子与买卖信号 → handlebar 按调仓日执行交易。
母版路径：本 skill 目录下的 scripts/daily-factors-backtest.py（相对 SKILL.md 所在目录）
适用场景
| 适合 | 不适合 |
|---|---|
| 日频截面因子选股（Barra 风格处理） | Tick/分钟高频 |
| 固定持仓数 Top-N 等权调仓 | 期货开平仓（qmt-future-trade，规划中） |
| 研报因子复现、上下影线/价值/动量等 | 目标持仓型期货实盘（qmt-live-strategy-template，规划中） |
| 申万行业 + 市值中性化 | 仅要信号推送（qmt-live-signal-feishu，规划中） |
母版架构（必须理解再改）
daily-factors-backtest.py
├── 文件头 # coding:gbk + 策略说明 docstring
├── 因子函数库 ← 【主要替换区】factor_xxx + 中性化/去极值
├── init(C) ← 【配置区】回测区间、股票池、资金、因子参数
├── after_init(C) ← 【信号区】拉数据 → 算因子 → 过滤 → 生成 g.buy/sell_signals
├── handlebar(C) ← 【执行区】调仓日卖出/买入（通常保留）
├── 交易执行函数 ← 通常原样保留
└── 辅助工具函数 ← 通常原样保留（IPO/ST/涨跌停/财务宽表）
各段职责
| 段 | 函数/变量 | 做什么 |
|---|---|---|
| 全局状态 | g = G() |
跨函数共享参数、信号矩阵、持仓 |
| 因子库 | factor_ubl(...) |
输入 OHLCV/市值等宽表，输出因子 DataFrame（index=日期, columns=股票） |
| 因子后处理 | filter_extreme_mad_df / neutralize_by_market_cap / neutralize_by_industry_zscore / cross_section_zscore |
Barra 风格流水线，按研报需求保留或删减 |
| 初始化 | init(C) |
设 g.start_date/g.end_date、g.stock_pool、g.max_positions、g.rebalance_days 等；预定义 g.buy_signals/g.sell_signals 防空矩阵 |
| 预计算 | after_init(C) |
一次性拉全区间行情+财务 → 算因子 → IPO/ST/停牌过滤 → 截面排名 → shift(1) 生成 T+1 信号 |
| 执行 | handlebar(C) |
每 g.rebalance_days 个交易日调仓：先卖后买，开盘价成交 |
| 交易 | execute_sell/buy_signals |
涨停不买、跌停不卖；科创板 200 股、其余 100 股整数倍 |
| 辅助 | get_ipo_mask / get_st_mask / get_financial_wide_table |
上市满 120 天、ST 区间、财务字段宽表 |
信号时序（防未来函数）
df_rank = df_factor_filtered.rank(axis=1, ascending=g.rank_ascending)
df_is_top_n = df_rank <= g.max_positions
g.buy_signals = df_is_top_n.shift(1).fillna(False) # T 日因子 → T+1 日买入
g.sell_signals = ~g.buy_signals
禁止去掉 .shift(1)，除非用户明确要求当日收盘调仓且接受前视偏差。
Agent 工作流
1. 解读策略输入
用户可能提供：文字描述、研报 PDF、截图、已有因子公式。提取并输出 策略规格表（生成前给用户确认）：
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 278 lines · 96 tokens per session scan A 1e6059072ae4
qmt-inner-backtest is a skill published in the GitHub repository dfkai/xtquantai (161 stars, last pushed 2mo ago), licensed MIT. It adds 96 tokens to every session and 3,719 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
skillshare-changelog
Generate CHANGELOG.md entry from recent commits in conventional format. Also syncs the website changelog page. Use this skill whenever the user asks to: generate a changelog, document what changed between tags, or create a new CHANGELOG entry. If you see requests like "write the changelog for v0.17", "what changed…
skillshare-update-docs
Update website docs to match recent code changes, cross-validating every flag against source. Use this skill whenever the user asks to: update documentation, sync docs with code, document a new flag or command, fix stale docs, or update the README. This skill covers all website/docs/ categories (commands, reference…
weekly-crm-report
Cleans a weekly CRM export and produces a regional sales summary. Activates when the user asks to clean a CRM export, deduplicate sales rows, calculate regional totals, or generate a weekly sales report from a CSV.
cloudbase-document-database-in-wechat-miniprogram
Use CloudBase document database WeChat MiniProgram SDK to query, create, update, and delete data. Supports complex queries, pagination, aggregation, and geolocation queries.
ulw-research
Team-first maximum-saturation research orchestration. ALWAYS asks which final format to render (PDF+DOCX default), then stands up a max-size cooperating team (teamcreate): one member per axis plus skeptic/red-team members for ultradebate/hyperdebate cross-critique, explore/librarian lanes, live journaling, an EXPAND…
dashscope
DashScope (Alibaba Cloud Bailian / 阿里云百炼) integration — image generation (qwen-image-2.0-pro), text-to-speech (qwen3-tts-flash), and ASR with word-level timestamps (qwen3-asr-flash-filetrans). Use when generating images via Qwen-Image, narrating via Qwen-TTS, or transcribing with word-level timestamps via Qwen-ASR.