Introduction

Under DSH’s “everything is a plugin” philosophy, many capabilities are distilled into installable and reusable skill packages. If you wish to perform A-share stock selection or position diagnosis within a session, the common approach is to concatenate data scripts, filtering rules, and risk warnings together. dsh-a-stock-select consolidates this process into a single skill. Below, I will introduce its positioning, installation, and usage.

What is This

dsh-a-stock-select is a DSH skill plugin developed based on a-stock-data (A-share data source). It embeds and merges all 47 data endpoints from a-stock-data V3.6.1, overlaying four major strategy filtering logic, position diagnosis workflows, and mandatory risk control disciplines, to create a self-contained SKILL.md—zero external dependencies, ready to use out of the box; meanwhile, it is packaged as an npm plugin, supporting one-click installation via the dsh command.

Verified information includes:

  • GitHub repository address: https://github.com/ct188579/dsh-a-stock-select
  • License: MIT
  • Node.js version requirement: >=18
  • Installation command: dsh plugin --profile web add dsh-a-stock-select

Core Features

Below are the verified functional points.

Four Stock Selection Strategies

The plugin provides four strategy timeframes:

  • Ultra-short term: 1–3 trading days
  • Short term: 1–4 weeks
  • Medium term: 1–6 months
  • Long term: >6 months

Each strategy provides a “fetch -> filter -> output” pipeline.

Position Diagnosis

Position diagnosis first performs a 7-step real data verification, then provides conclusions such as Hold / Reduce Position / Close Position, along with support levels, resistance levels, stop-loss levels, and triggering conditions.

Built-in Data Endpoints

The plugin includes 47 data endpoints from a-stock-data V3.6.1 for running code.

Mandatory Risk Control

Every strategy has mandatory stop-loss discipline. The plugin does not recommend ST / key monitoring pool / limit-down pool targets, and every output forcibly includes risk warnings and disclaimers.

Installation and Activation

Before one-click installation of DSH, your local machine must have DSH and pnpm installed. Execute the following command to install the plugin to the web profile:

dsh plugin --profile web add dsh-a-stock-select

After installation, restart DSH, and the skill a-stock-select will be registered to that profile.

Manual installation requires a Python 3 environment and the installation of these dependencies:

pip install requests pandas stockstats mootdx

Since the plugin runs with the permissions of the current DSH process, you should check the source code and license before installation.

Typical Usage

You can trigger the skill using natural language, for example:

  • “Recommend a few short-term stocks”
  • “Help me select stocks for the long term”
  • “I hold Datang Power, cost X yuan, what should I do?”
  • “What are the directions for next week?”

These examples correspond to short-term stock selection, long-term stock selection, position diagnosis, and directional inquiry.

Applicable Scenarios and Notes

This plugin is suitable for individuals who wish to use fixed A-share stock selection and position diagnosis workflows within DSH sessions, particularly developers who need to run strategy filtering, data verification, and risk control prompts within a single skill.

Please note before use:

  • The stock market involves risks; please be cautious when entering the market.
  • Plugin output is for research reference only and does not constitute any investment advice.
  • Operate at your own risk for profits and losses; losses are unrelated to the author, this tool, or the data sources.
  • Data may be delayed, missing, or incorrect.
  • Interfaces may become unreachable due to risk control / network issues, with a built-in fallback chain.
  • Please check the source code and license before installation; the plugin runs with the permissions of the current DSH process.

Links

The DSH community directory is an independent site with no official affiliation with DeepSeek or Hypersphere. The URL of the directory page was not obtained in this article; the GitHub repository address is as follows:

  • https://github.com/ct188579/dsh-a-stock-select