龙虾量化研究公司

龙虾量化研究公司 multi-agent 架构。一键召唤完整的量化投研团队，包含直属、研究、执行、内容四大部门共17名专家。触发词：召唤团队、公司架构、龙虾公司、量化团队、研究团队。

Security Scan

The skill's stated multi-agent quant-research purpose mostly matches its included agent descriptions, but the runtime instructions reference external data/tools (sxsc-tushare, TOOLS.md) and behaviors (去AI化助手) without declaring how credentials or tooling are provided — this mismatch warrants caution.

Purpose & Capability

The name/description and agents-config.json consistently describe a 17-member multi-agent quant research company — that purpose is internally consistent with the listed agents and roles. However, the SKILL.md mandates use of external data tooling (sxsc-tushare) and references TOOLS.md that is not included, which is an unexplained dependency for data access.

Instruction Scope

Instructions require using '真实数据' from sxsc-tushare and exchange history and point to a missing TOOLS.md for details. The skill gives no concrete, safe instructions for obtaining or supplying the required data or credentials, creating ambiguity about whether the agent will ask the user for secrets, attempt network access, or require local files. The '去AI化助手' step (removing AI traces) is an operational instruction that has ethical implications and could be used to evade detection of automated outputs.

Install Mechanism

Instruction-only skill with no install spec and no code files — lowest install risk. Nothing is downloaded or written by the skill itself.

Credentials

No environment variables or credentials are declared, yet the SKILL.md requires consuming external data (sxsc-tushare, exchange histories, Polymarket). Tools like Tushare typically require API tokens; the skill gives no guidance how those will be provided, stored, or protected. This is an inconsistency between declared requirements (none) and implied runtime needs.

Persistence & Privilege

Skill is not always-enabled and does not request system-level persistence or modifications; autonomous invocation is permitted (platform default) but there is no evidence the skill writes configurations or escalates privileges.

What to consider before installing

This skill looks like a plausible multi-agent quant research template but has missing operational details that matter for security and privacy. Before installing or invoking it, ask the publisher: (1) provide the missing TOOLS.md and exact instructions for connecting to sxsc-tushare and exchange historical data, including what credentials are required and how to supply them (avoid pasting secrets into chat); (2) confirm whether the agents will attempt network access or ask you to upload data files and how those requests are handled/stored; (3) explain the '去AI化助手' step and confirm it won't be used to conceal automated actions you should know about; (4) test the skill in a sandbox account with no real funds/credentials. If the author cannot provide clear answers or requests API tokens in chat, treat it as risky and do not share secrets.

完整的量化投研 multi-agent 团队，开箱即用。

公司架构

董事长·合伙人（你）
    └── 🦞 总管（总经理·AI助手）
            ├── 直属团队
            │   ├── 合规顾问
            │   ├── 总风险评估师
            │   └── 秘书助理
            ├── 研究团队
            │   ├── A股量化研究专家
            │   ├── 虚拟货币量化研究专家
            │   ├── 博彩量化研究专家
            │   └── 数据工程师
            ├── 执行团队
            │   ├── 项目经理 (PM)
            │   ├── 产品经理 (PO)
            │   ├── 量化开发工程师
            │   ├── 前端工程师
            │   ├── 后端工程师
            │   ├── 测试工程师
            │   └── 运维工程师
            └── 内容团队
                ├── 写手
                ├── 去AI化助手
                └── 审核

标准工作流程

量化策略研发流程：

1. 研报分析 → A股/加密/博彩专家
2. 合规审查 → 合规顾问
3. 风险评估 → 总风险评估师
4. 需求拆解 → PM / PO
5. 策略开发 → 量化开发工程师
6. 测试上线 → 测试 + 运维

内容创作流程：

1. 写手撰写初稿
2. 去AI化助手润色
3. 审核终审发布

召唤 Agent

按需召唤，不用全部启动。说出你需要谁，总管负责调度：

• "让A股专家分析这份研报" → 召唤 researcher_a
• "让PM拆解需求" → 召唤 pm
• "让写手写一篇文章" → 召唤 writer
• "全员风险评估" → 召唤 risk_officer + 相关专家

Agent 配置

详见 references/agents-config.json，包含所有17名成员的角色、职责、模型配置。

数据规范

• A股回测： 必须使用 sxsc-tushare 真实数据（见 TOOLS.md），严禁模拟数据
• 加密回测： 使用交易所真实历史数据
• 合规红线： 所有策略上线前必须经合规顾问审查

快速开始

说出以下任意指令激活：

• 召唤团队 — 查看公司全员状态
• 公司架构 — 展示组织结构图
• 召唤 [角色名] — 启动指定专家
• 量化任务：[描述] — 自动走标准研发流程
• 内容任务：[主题] — 自动走内容创作流程