myc0576/SmartMoney-Cub

Name: myc0576/SmartMoney-Cub
Author: myc0576

Plugin插件 ⭐ 26 MIT Data & Analysis数据与分析

smartmoney-cub-harness is a local-first, agent-agnostic trading journal and review harness: read-only over markets and execution, writable over your own journal. It turns an external caller's offline run into portable, reviewable artifacts without taking trading authority.

catalog descriptioncatalog 简介 / catalog description：Read-only trading journal and review harness: Jev typed judgments, agent integration, and a reproducible finance benchmark. No orders, no advice.

Project Overview项目介绍

Edited by Zhiwei Agent (zhiwei)由知微Agent(zhiwei)编辑

SmartMoney-Cub is a local-first, agent-agnostic trading journal and review harness that works with any AI agent including DeepSeek Harness. It captures offline decision runs from external agents or CLIs and turns them into portable, reviewable artifacts without taking any trading authority on your account. To install it for first use, you simply clone the repository to your local machine, install the package with pip in editable development mode, then run the provided toy workflow sequence to test its full core functionality offline. This gives you a quick sense of how the tool works before you use it for your own trading data.

The standard workflow starts with capturing a completed run from your connected AI agent, then generates a cryptographically sealed run envelope and evidence pack that holds all relevant decision context for your review. It then runs a deterministic replay of the decision against delayed D1 or D3 outcome data, evaluates the candidate strategy, and requires an explicit human approval gate before any updated rule can move to the next planning cycle. This tool is designed for individual traders and quantitative researchers who want to backtest, review, and structurally log AI-assisted generated trading strategies over extended periods of time.

This project requires Python 3.10 or newer and is released under the permissive MIT open source license. Its core runs entirely offline, has no embedded LLM, no broker connection, and does not support automatic trading of any kind. The project explicitly states it is not financial advice, and all user trading journal data is stored locally or in the user’s own tenant store, never in the public project repository. When running untrusted code, you should always use an OS or container sandbox, as the built-in sandbox flag only sets an output namespace and does not isolate processes.

SmartMoney-Cub 是一个优先本地运行、与 AI 代理无关的交易日志复盘工具框架，核心功能是将外部 AI 代理的离线交易决策运行转换为可移植、可复盘的标准化工件，本身不具备任何交易执行权限。它支持包括 DSH 在内的各类 AI 代理集成，可通过 CLI 命令快速完成安装和离线示例运行，帮助用户梳理和复盘个人的交易决策逻辑。

典型工作流从捕获 AI 代理的决策运行开始，生成标准化运行信封和证据包，之后通过确定性回放对决策结果进行回溯评估，整个流程要求人工介入审核，只有通过人工审核的决策规则才能进入下一轮计划。它面向需要对 AI 辅助生成的交易策略进行回测、复盘和结构化记录的个人交易者、量化研究者使用。

本项目基于 Python 3.10+ 开发，采用 MIT 许可证开源，核心完全离线运行，不内嵌大语言模型，也不连接券商接口，不支持自动交易。项目明确声明不提供任何投资建议，仅用于研究、日志记录和工作流设计，用户的交易日志存储在本地或用户自己的租户存储中，不会上传到项目仓库。

View on GitHub在 GitHub 查看

Send to your DSH发给你的 DSH 安装

请帮我安装这个 DSH 插件。安装前先完成【兼容性检查 + 安全性检查】，检查通过再动手。

插件：SmartMoney-Cub（myc0576/SmartMoney-Cub）
仓库：https://github.com/myc0576/SmartMoney-Cub
本站详情页：https://www.yhbd.top/plugins/myc0576-smartmoney-cub/
本站登记：类型 plugin · 归类 多平台兼容工具（非 DSH 原生） · 许可证 MIT · ⭐ 26 · 最近提交 2026-09-25 · 主语言 Python · 未检测到 DSH 插件清单

按下面顺序执行，每步先把结论告诉我，再进入下一步：

【1 兼容性检查】
① 我这边：DSH 版本、Node 版本、操作系统、当前 profile（web / desktop）。
② 读它的 README、package.json、插件 manifest，列出它要求的 DSH 版本 / Node 版本 / 操作系统 / 外部依赖 / 需要另外先装的运行时。
③ 逐条比对，结论只写「满足 / 不满足 / 未知」三种；不满足的给出可行替代方案。
④ 检查是否和我已装的插件冲突：命令名重复、skill / tool 重名、端口占用、重复注册的 MCP server。

【2 安全性检查】
① 仓库可信度：和上面「本站登记」是否一致；star / fork 数、创建时间、最近提交，是否归档或长期停更。
② 安装脚本：逐行看 package.json 的 preinstall / install / postinstall，以及 install.sh、setup.ps1 之类脚本。出现 curl|bash、下载后直接执行、混淆代码、访问与插件功能无关的域名，立刻停下来告诉我，不要继续装。
③ 依赖：列出新增依赖，标出无人维护、或与知名包拼写近似的可疑包（typosquatting）。
④ 权限与副作用：它会读写哪些目录、访问哪些域名、需要哪些 DSH 权限（filesystem / network / shell / clipboard 等），以及怎么卸载和回滚。
⑤ 如果它要求 sudo / 管理员权限，或权限明显超出功能所需，先停下来问我。

【3 安装】
上面两步没有「不满足」和「高危项」时才执行；用官方推荐方式安装，不要自行提权。

【4 汇报】
用表格输出：检查项 / 结论 / 依据 / 是否需要我决策。拿不准的一律写「未知」并说明要我怎么确认——不要猜，也不要替我决定。

Send this message to DSH in your current session: it verifies compatibility and security first (answering met / not met / unknown item by item) and only installs once everything checks out — it will stop and ask you if it finds a high-risk item. The box scrolls; the copy is the full prompt. CLI install commands may not be accurate across systems, so DSH is the safer route.把上面这条消息直接发给当前会话里的 DSH：它会先核对兼容性与安全性（逐条给「满足 / 不满足 / 未知」），确认没问题再安装，有高危项会停下来问你。框内可滚动，复制到的是完整提示词；安装命令不一定准确，发给 DSH 更稳。

Pre-install check安装前体检Compatibility · Security兼容性 · 安全性 1 warning1 项注意

26 stars - an early-stage project星标 26，属于早期项目
No DSH plugin manifest detected - it may only carry the dsh-plugin topic, so the install method must be confirmed on the spot未检测到 DSH 插件清单：可能只是打了 dsh-plugin 话题，安装方式要现场确认
Not DSH-native: a multi-platform tool that may require Node / Electron or another runtime first非 DSH 原生，是多平台兼容工具：可能要先装 Node / Electron 等运行时

DSH walks through these 9 checksDSH 会逐条核对这 9 项

Compatibility兼容性

DSH, Node, OS and profile requirementsDSH 版本 / Node 版本 / 操作系统 / profile 是否满足要求
External dependencies and runtimes (Electron / Python / Docker, ...)外部依赖与运行时（Electron / Python / Docker 等）是否齐备
Conflicts with installed plugins: command names, skill / tool names, ports, duplicate MCP registration与已装插件是否冲突：命令名、skill / tool 重名、端口占用、重复 MCP 注册

Security安全性

Repo matches the facts registered here; archived or abandoned?仓库是否与页面登记一致，是否归档或长期停更
Safety of preinstall / install / postinstall and install.sh / setup.ps1preinstall / install / postinstall 与 install.sh、setup.ps1 是否安全
curl|bash, download-then-execute, obfuscation, unrelated domains → stop immediatelycurl|bash、下载即执行、混淆代码、无关域名 → 立刻停止
Typosquatting or unmaintained packages among the new dependencies新增依赖里有没有 typosquatting 或无人维护的包
Requested permissions vs. what the feature actually needs申请了哪些权限、是否超出功能所需（filesystem / network / shell / clipboard）
Any sudo / admin requirement, plus uninstall and rollback是否要求 sudo / 管理员权限，以及卸载与回滚方式

Anything uncertain must be marked unknown with a note on how to confirm it. This site's signal screen is a static snapshot, not a security audit.拿不准的必须标「未知」并说明要我怎么确认。本站的信号筛查是静态快照，不能替代安全审计。

Or use CLI install (for developers)或使用命令行安装（适合开发者）

CLI Install命令行安装

dsh plugin --profile web add github:myc0576/SmartMoney-Cub

把 myc0576/SmartMoney-Cub 加入你的 DSH 配置（web profile）即可启用。

READMEREADME

35.7 KB · 缓存于 2026-09-25 22:09:09 · 35.7 KB

smartmoney-cub-harness

SmartMoney-Cub bilingual cover

smartmoney-cub-harness is a local-first, agent-agnostic trading journal and review harness: read-only over markets and execution, writable over your own journal. It turns an external caller's offline run into portable, reviewable artifacts without taking trading authority.

External Agent or CLI caller → Run Envelope → frozen Benchmark/Evidence Pack → deterministic replay → explicit human promotion gate.

Finance-JEV Benchmark Hero

Jev Reasoning Layer & Four-Track Financial Benchmark

SmartMoney-Cub supports Jev (TypeSafe | OpenRouter) as an optional typed-judgment layer with two pluggable backends: TypeSafe direct and OpenRouter. Jev evaluates only structured noul, choice, and score judgments, while all arithmetic, date comparisons, and strict temporal boundary validation (available_at <= decision_time) remain enforced in deterministic Python code.

The repository ships finance-jev-v1, a frozen offline evaluation suite containing 240 cases across four tracks (trading-review, financial-filings, industry-events, macro-policy). Following full answerability auditing and the elimination of input label leakage, cases present realistic evidence narratives (post-trade logs, disclosure excerpts, wire dispatches, central bank communiques) evaluated against strictly typed questions without answer leakage. In the published reference run (assets/benchmark/run.json), the deterministic rule baseline achieves 83.33% overall accuracy (95% Wilson confidence interval [80.92%, 85.49%]) and a macro F1 of 0.7792 across all 240 cases (1,020 evaluated items). The live TypeSafe Jev backend (typesafe_direct, evaluated against the real API resolving to model jev-1.13.0) achieves 78.43% overall accuracy (95% Wilson confidence interval [75.80%, 80.85%]) and a macro F1 of 0.7319 with a median latency of 1034 ms and superior probabilistic calibration (ECE of 0.1464 vs 0.1667). On financial-filings, Jev achieves 77.00% accuracy (F1: 0.6990) outperforming the baseline (73.33%), while achieving 93.33% accuracy (F1: 0.9215) on industry-events and 74.44% (F1: 0.7148) on macro-policy. These figures reflect empirical performance on this frozen toy suite and do not imply generalization to production market regimes. The OpenRouter backend (openrouter_jev) remains reported as not_run due to unconfigured credentials.

Showing the opening section of the README — the full document lives in the repository以上为 README 开头摘要，完整文档在仓库内 · View the full README on GitHub →在 GitHub 查看完整 README →

渲染基于仓库 myc0576/SmartMoney-Cub 的 README.md。在 GitHub 查看原文 →