AI Agent Hub
Back to skills
Tonghuashun Finance Data Query icon

Tonghuashun Finance Data Query

Professional Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please install @user_c163d6d4/ifind-finance-data2 according to https://skillhub.cn/install/skillhub.md.

About this skill

What problem it solves

Financial data queries often span multiple endpoints: stocks, funds, macro indicators, news and notices, bonds, HK/US stocks, and indices. When engineers or agents manually assemble requests, handle tokens, and compose multi-entity or multi-metric queries, they can easily run into parameter format, tool naming, and concurrency issues. This skill wraps the Tonghuashun finance data MCP service so scattered financial lookups share one stable interface, reducing repeated URL and token handling.

How it works

The core function is call(server_type, tool_name, params), routing requests by service type: stock, fund, edb, news, bond, global_stock, and index. It supports Node.js call-node.js and Python call.py; when the Python environment is unclear, prefer Node.js because it has no extra dependencies.

Core capabilities include:
- Stocks: smart stock selection, basic data, financials, quotes, shareholders, risk indicators, ESG ratings, and major events
- Funds: search, profiles, performance, holdings, holder structure, fund company data, and financials
- Macro and industry: use search_edb to find indicators, then get_edb_data to retrieve values
- News and notices: semantic search for news, company notices, and trending events
- Bonds: basic info, market data, issuer financials, and special indicators
- HK/US stocks: selection, profiles, quotes, financials, IPO/buyback/dividend events
- Indices and sectors: index quotes, sector performance, and constituent metrics

Key steps:
1. Choose the server_type based on the data target, such as stock for equities or edb for macro indicators
2. Use documented tool_name values first instead of starting with list_tools
3. Put conditions in query or keyword, using YYYY-MM-DD for dates
4. Merge multi-entity and multi-metric requests, keeping entities and metrics within 5
5. Check the returned ok field to confirm success

Boundaries and cautions

The skill requires a valid auth_token, typically stored in config.json or mcp_config.json, and needs either Node.js or Python. Free, personal, and enterprise tiers allow up to 2, 5, and 10 concurrent requests per second; default to the free limit when the tier is unknown. Avoid overly broad sector or time ranges to prevent truncation, and keep trending-news parameters narrow because too many constraints can return no results. Use list_tools only when a tool is missing, renamed, or inconsistent with the documentation. After each query, clean up any temporary scripts generated for the task.

Use Cases

  • When building stock research answers, query electronic-industry stocks with market cap over 10 billion.
  • In daily research scripts, batch-fetch net profit growth, ROE, and ROA for multiple listed companies.
  • When uncertain about PV supply-chain indicators, search first and then retrieve production data.
  • For notice monitoring, search company, date range, and topic to retrieve relevant news or notice snippets.

Best For

  • Quant analysts doing A-share single-stock research who need quick access to financials, quotes, shareholders, ESG, and event data.
  • Investment-research engineers building automated daily reports that combine macro indicators, industry metrics, and stock/fund data.
  • Application developers building finance Q&A agents that convert natural-language stock/fund selection into MCP tool calls.
  • Research operations staff who need date- and topic-based retrieval of notices, financial news, and trending events.