Mobile Tonghuashun ADB Trading
Paste the following prompt into your AI chat to install this skill:
Please install @user_d5f611dc/ths-adb-trader using https://skillhub.cn/install/skillhub.md.
About this skill
Problem
When managing a Tonghuashun account on a phone screen, an agent has to repeatedly tap through UI flows for account checks, order review, cancellation, buy, and sell. This skill wraps Android phone + Tonghuashun APP + ADB UI automation into a local MCP service, so engineers can drive mobile trading workflows with natural language or tool calls.
How It Works and Limits
It exposes 16 tools on a localhost:19090 MCP endpoint, with typical paths:
- Read-only checks: get_account, get_positions, get_orders, get_trade_mode, screenshot
- Controlled writes: buy_stock, sell_stock, cancel_order, set_trade_mode
Key constraints: default to simulate mode; confirm trade_mode before buy, sell, or cancel; verify submitted orders with get_orders; on failure use screenshot and dump_ui instead of blind retries. It does not pick stocks, perform fundamental or technical analysis, predict prices, or replace a strategy system. Real orders still depend on phone connectivity, APP version, network state, and UI changes, and trading P&L is the user’s responsibility.
Use Cases
- When asked to check account balance and positions, use `get_account` and `get_positions` on Tonghuashun.
- When instructed to buy 600519 at market, confirm `trade_mode`, run `buy_stock`, and verify the order.
- When debugging APP freezes or broken UI flows, capture `screenshot` and export `dump_ui` state.
- When testing cancellation in simulate mode, use `get_orders` to locate the order, then call `cancel_order`.
Best For
- Quant engineers maintaining mobile ADB automation who need account, order, and buy/sell calls from a PC.
- MCP agent tooling developers who want to turn mobile UI steps into controlled tool calls.
- Trade support staff testing pre-trade checks who need simulate-mode account, order, and cancel validation.
- Strategy system owners managing real-trading risk who need default simulate mode and explicit real-mode confirmation.
Related Skills
Switch AI from a polite executor into a skeptical thinking partner, using assumption checks, pre-mortems, and evidence gates to expose weak ideas early.
Applies research discipline, data lookup, evidence grading, red-team checks, and structured reporting to analyze AI company investment logic.
Uses Qichacha or Tianyancha MCP data and public information to score corporate credit across 12 dimensions and output a Markdown risk report.
Calculates personal injury compensation via CLI and outputs case summary, standards, line items, sources, and risk notes.