Investment Workflow

v1.2.0

[何时使用]当用户需要投研分析时;当用户说'分析这个标的'、'现在什么值得买'、'对 XX 行业怎么看'、'这个热点有什么影响'、'开会讨论投资'时触发。场景驱动的投研全流程,覆盖 6 个场景。

0· 98·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for lj22503/investment-workflow.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Investment Workflow" (lj22503/investment-workflow) from ClawHub.
Skill page: https://clawhub.ai/lj22503/investment-workflow
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install investment-workflow

ClawHub CLI

Package manager switcher

npx clawhub@latest install investment-workflow
Security Scan
Capability signals
Requires sensitive credentials
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
Name/description (investment research workflow) align with the instructions and referenced dependencies: the SKILL.md consistently describes data-query, market-scan, industry-rank, etc., and references existing data tooling (data_layer, mcp-aktools). There are no unexplained environment variables, unrelated binaries, or requests for cloud credentials that would be out of scope for an investment workflow.
Instruction Scope
The runtime instructions ask the agent to call local data providers (data_layer, mcp-aktools), use cached local data on failure, and route final content to an expression-layer for formatting/delivery. These are coherent for the skill's purpose, but the skill grants the agent Exec/Read/Write permissions (allowed-tools) and references a home-path binary (~/.local/bin/mcp-aktools) and a local data_layer directory — you should confirm those tools and their data sources are trusted and that expression-layer routing (which may deliver output externally) is acceptable.
Install Mechanism
This is an instruction-only skill with no install spec and no code files. That minimizes risk from arbitrary downloads or archive extraction. The clawhub metadata lists required components (data_layer v2.2.0, mcp-aktools) but installation is external to this skill.
Credentials
The skill requests no environment variables, credentials, or config paths beyond referencing expected local tool paths. There are no unrelated secret requests. Its need to read local data_layer caches is proportional to the stated data-query purpose.
Persistence & Privilege
always:false and default autonomous-invocation behavior are appropriate. The skill does permit Exec/Read/Write when running, but it does not request permanent presence or modification of other skills' configs. Review expression-layer to understand any external delivery side effects.
Assessment
This skill appears coherent for investment research: it relies on local data tooling (data_layer, mcp-aktools) and formats output via an expression-layer. Before installing or enabling it, verify: (1) the local data tools (data_layer and ~/.local/bin/mcp-aktools) are the expected trusted implementations and their data sources (AKShare) are acceptable; (2) the expression-layer it calls (../expression-layer) does not automatically publish reports to external services you don't want (e.g., WeChat, email, or third-party servers); (3) you are comfortable the agent will be allowed to Exec/Read/Write local data_layer caches — run in a limited environment if you want to restrict filesystem/network exposure. No credentials are requested by this skill, which reduces risk.

Like a lobster shell, security has layers — review code before you run it.

latestvk97dcxcf91d439w0es89mv9dbh85kjry
98downloads
0stars
3versions
Updated 1d ago
v1.2.0
MIT-0

investment-workflow: 投资工作流 🎯

📋 功能描述

帮助用户系统化执行投研全流程。不是一套固定流程,是 6 个用户场景 × 共享 Skill 模块的排列组合。

适用场景:

  • 买股票 / 投行业 / 扫描推荐 / 行业看法 / 热点分析 / 开会讨论

边界条件:

  • 不替代深度基本面研究
  • 输出为 Markdown 报告,需配合 data_layer / mcp-aktools 获取真实数据
  • 场景识别依赖用户输入关键词
  • 模糊输入处理:用户输入过于模糊(如"最近怎么样?")时,先询问澄清(行业/事件/标的),不强行触发完整工作流

🔄 6 个核心场景

场景触发词调用步骤输出
买股票"我要买股票"、"分析 XX"data-query → stock-research → decision-integrateMarkdown 决策报告
投行业"我要投行业"、"XX 行业值得投吗"industry-rank → data-query → multi-view → decision-integrateMarkdown 行业报告
扫描推荐"现在什么值得买?"market-scan → industry-rank → data-query → decision-integrateTop 3 标的 + 优先级
行业看法"对 XX 行业怎么看"industry-rank → deep-think → plain-explainMarkdown 散文
热点分析"这个热点有什么影响"market-scan → industry-rank → multi-view → plain-explainMarkdown 影响分析
开会讨论"开会讨论"multi-view → decision-integrateMarkdown 会议纪要

详细共享 Skill 说明 → references/shared-skills.md 数据层集成说明 → references/data-layer-integration.md


🔄 核心执行流程(6 阶段)

原则:按场景调用 2-4 个阶段,非必须不跑全 6 步。

阶段①:场景识别

动作:从用户输入提取关键词,匹配 6 个核心场景之一。 输出:场景名称 + 触发步骤列表。 降级:若输入模糊(如"最近怎么样?"),先询问澄清,不强行触发。

阶段②:数据查询(market-scan / data-query / industry-rank)

动作:调用 data_layer / mcp-aktools 获取实时数据。 指令

  • 优先调用 data_layer(带缓存)。
  • 失败时降级到本地缓存或明确告知"数据获取失败"。
  • 标注数据来源与时间:[数据:指标 | 来源:数据源 | 时间:时间]

阶段③:深度分析(stock-research / multi-view / deep-think)

动作:执行降秩分析、追本分析或多视角验证。 指令

  • 必须保留推演链(表象→机理→原理→公理)。
  • 附 ASCII 关系图辅助理解。
  • 合规转化:若用户请求"帮我买 XX"或"推荐代码",不直接拒绝,转化为"交易成本分析"或"筛选方法论"投教内容。

阶段④:逻辑验证(decision-checklist / mental-models)

动作:使用芒格思维模型/检查清单验证分析逻辑。 输出:逻辑漏洞清单 + 修正建议。

阶段⑤:决策整合(decision-integrate)⭐ 核心

动作:综合所有分析,给出明确建议。 指令

  • 必须输出买入 / 卖出 / 持有 / 观望 之一。
  • 必须标注:置信度(高/中/低)及理由。
  • 若数据不足,明确说明"因 XX 数据缺失,置信度低"。
  • 禁止:只罗列数据表格,不给结论。

阶段⑥:多形式输出(调用表达层)⭐ 核心

动作:将决策报告传递给 expression-layer 进行格式化。 指令

  • 调用 expression-layer,传入 content + intent(plain/writes/card/wechat)。
  • 禁止:直接输出原始 Markdown,必须经表达层路由。
  • 输出形式:Markdown 报告 / PNG 卡片 / HTML 演讲 / 公众号推文。

⚠️ 常见错误

错误 1:线性执行所有阶段

问题:
• 用户只问"XX 怎么看",却跑完 5 个阶段
• 输出冗长,用户找不到重点

解决:
✓ 严格按场景定义调用 2-4 个共享 Skill
✓ 输出聚焦场景核心问题

错误 2:忽略数据层降级

问题:
• mcp-aktools 或 data_layer 调用失败
• 报告数据为空或报错

解决:
✓ 优先调用 data_layer(带缓存)
✓ 失败时降级到本地缓存或明确告知"数据获取失败"
✓ 标注数据来源与时间

错误 3:推理过程缺失

问题:
• 直接给结论,用户不知道"怎么想到的"
• 缺乏降秩/追本/验证过程

解决:
✓ 每个结论必须保留推演链
✓ 使用 [数据:指标 | 来源:数据源 | 时间:时间] 格式
✓ 附 ASCII 关系图辅助理解

错误 4:数据罗列无结论

问题:
• 阶段③只输出数据表格,阶段⑤未给出明确建议
• 用户得不到 actionable 的建议

解决:
✓ 阶段⑤必须综合数据给出"买/卖/持有"建议及置信度
✓ 使用 decision-integrate 共享 Skill 整合所有分析

错误 5:合规拦截过度

问题:
• 遇到模糊或交易指令,直接拒绝
• 未提供替代方案(如"扫描当前强势板块")

解决:
✓ 执行"合规转化"协议:拒绝 -> 转化为分析/投教
✓ 例:"无法直接交易" → "分析交易成本/时机"

错误 6:表达层未调用

问题:
• 阶段⑥未调用 expression-layer
• 输出格式不统一(有时表格,有时纯文本)

解决:
✓ 阶段⑥必须调用 expression-layer,传入 content + intent
✓ 让表达层统一路由至 Markdown/PNG/HTML

🧪 使用示例

输入:

分析消费 ETF 是否值得投?

预期输出:

  • 识别场景:投行业
  • 调用:industry-rank → data-query → multi-view → decision-integrate
  • 输出:Markdown 报告(含降秩分析、数据验证、圆桌讨论、决策建议)

输入:

现在什么值得买?

预期输出:

  • 识别场景:扫描推荐
  • 调用:market-scan → industry-rank → data-query → decision-integrate
  • 输出:Top 3 标的 + 优先级 + 逻辑

🔧 故障排查

问题检查项
不触发description 是否包含触发词?用户输入是否匹配场景?
数据为空data_layer 是否安装?mcp-aktools 是否运行?缓存是否过期?
输出过长是否跨场景调用?检查场景定义,只调用必要步骤
推理缺失是否跳过降秩/追本步骤?检查共享 Skill 执行顺序

🔗 相关资源

  • 共享 Skill 文档:references/shared-skills.md
  • 数据层集成:references/data-layer-integration.md
  • 报告模板:templates/report-template.md
  • 表达层路由:../expression-layer/SKILL.md
  • 标准参考:docs/SKILL-STANDARD-v3.md

Comments

Loading comments...