← 返回
未分类 Key 中文

Finnhub Skill

Read-only market data skill for Finnhub. Use when the user wants stock, forex, crypto, company profile, candles/K-lines, news, earnings, or economic calendar...
仅读取 Finnhub 市场数据技能。仅在用户需要股票、外汇、加密货币、公司概况、K线/蜡烛图、新闻、财报或经济日历时使用。
samlin425 samlin425 来源
未分类 clawhub v0.2.0 1 版本 100000 Key: 需要
★ 0
Stars
📥 330
下载
💾 1
安装
1
版本
#latest

概述

Finnhub Skill

Use Finnhub as a read-only market data source.

Core Rules

  1. Treat this skill as data retrieval only.
  2. Never place trades or modify accounts through this skill.
  3. Require a valid FINNHUB_API_KEY before making live requests.
  4. If the API key is missing, explain how to configure it instead of guessing.
  5. Prefer concise outputs: summary first, raw fields second.

Supported Use Cases

  • Real-time quote lookup
  • K-line / candle retrieval
  • Company profile lookup
  • Symbol/company news lookup
  • Earnings calendar / earnings history lookup
  • Economic calendar lookup
  • Crypto and forex reference data (if supported by the user plan)
  • Structured daily stock news report (price + news + trader-style interpretation)

Configuration

Expect the user to provide:

export FINNHUB_API_KEY=your_api_key

Optional base URL override is supported only for the official Finnhub domain:

export FINNHUB_BASE_URL=https://finnhub.io/api/v1

Default base URL:

https://finnhub.io/api/v1

Security rule:

  • Only allow https://finnhub.io/...
  • Do not point this skill to arbitrary hosts or proxies that could capture the API key

Execution Layer

Use the bundled script:

python3 scripts/finnhub.py <command> [flags]

Supported commands:

  • quote --symbol AAPL [--raw]
  • candles --symbol AAPL --resolution D --from-ts 1711584000 --to-ts 1712188800 [--raw]
  • profile --symbol AAPL [--raw]
  • company-news --symbol AAPL --date-from 2026-03-01 --date-to 2026-03-30 [--raw]
  • market-news --category general [--raw]
  • earnings --date-from 2026-03-30 --date-to 2026-04-06 [--symbol AAPL] [--raw]
  • economic --date-from 2026-03-30 --date-to 2026-04-06 [--raw]

Default output is human-readable. Use --raw when the user explicitly wants JSON.

Recommended Request Pattern

Use the bundled Python script for live calls instead of rebuilding requests from scratch.

When constructing requests:

  • URL-encode symbol and other parameters
  • Validate time range inputs before sending
  • Keep requests read-only
  • If the user asks for a very broad request, narrow the scope first

Output Style

Quote

Return:

  • symbol
  • current price
  • absolute change
  • percent change
  • high / low / open / previous close
  • timestamp if available

Candles

Summarize:

  • symbol
  • resolution
  • start/end window
  • number of candles returned
  • latest OHLCV row

Only dump the full array when the user explicitly asks.

Company Profile

Return:

  • company name
  • ticker
  • exchange
  • currency
  • country
  • market cap
  • industry / IPO date if available
  • website

News

Return:

  • headline
  • source
  • published time
  • URL
  • short summary if available

Prefer 3-5 most relevant items unless the user asks for more.

Common Tasks

1. Real-time quote

Use Finnhub quote endpoint for stocks or supported symbols.

If the user gives a bare ticker like AAPL, use it directly.

If the user gives crypto/forex, confirm Finnhub symbol format if needed.

2. Candles / K-lines

Ask for or infer:

  • symbol
  • resolution (1, 5, 15, 30, 60, D, W, M)
  • from timestamp
  • to timestamp

If the user asks loosely (e.g. “last week”), convert it into a concrete range.

3. Company profile

Use the company profile endpoint when the user asks “what is this company”, “profile”, “market cap”, “which exchange”, etc.

4. News

Use company or market news endpoints depending on the request:

  • company-specific → company news
  • broad market / macro → market news

5. Earnings / calendar

Use earnings calendar for upcoming results and earnings history if the user asks what already happened.

6. Daily stock report

When the user asks for a daily report like:

  • “发我 TSLA 昨天的新闻总结”
  • “做一个 NVDA 昨日新闻日报”
  • “给我 AAPL 的价格+新闻日报”

Build the report in this order:

  1. quote / price summary first
  2. most important 3-5 news items
  3. market narrative and trader interpretation
  4. final rating

For the report format, read:

  • references/daily-report-template.md

If candle/volume endpoints are unavailable due to Finnhub plan limits:

  • still produce the report
  • explicitly say that detailed candles/volume are unavailable under current access
  • avoid inventing volume comparison or trend detail

Error Handling

If Finnhub returns auth or quota errors:

  • State the likely cause clearly
  • Do not fabricate fallback data
  • Suggest checking API key, plan limits, or symbol format

If symbol format is ambiguous:

  • Ask one clarifying question
  • Do not assume unsupported exchange suffixes

Reference File

For endpoint patterns and parameter hints, read:

  • references/api.md

版本历史

共 1 个版本

  • v0.2.0 当前
    2026-05-03 11:18 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

professional

A股量化 AkShare

mbpz
A股量化数据分析工具,基于AkShare库获取A股行情、财务数据、板块信息等。用于回答关于A股股票查询、行情数据、财务分析、选股等问题。
★ 186 📥 62,006
professional

All-Market Financial Data Hub

financial-ai-analyst
基于东方财富数据库,支持自然语言查询金融数据,覆盖A股、港股、美股、基金、债券等资产,提供实时行情、公司信息、估值、财务报表等,适用于投资研究、交易复盘、市场监控、行业分析、信用研究、财报审计、资产配置等场景,满足机构与个人需求。返回结果为
★ 122 📥 41,356
professional

Stock Analysis

udiedrichsen
{"answer":"基于雅虎财经数据,分析股票与加密货币。支持投资组合管理、自选股预警、股息分析、8维评分、热门趋势扫描及传闻/早期信号探测。适用于股票分析、持仓追踪、财报异动、加密监控、热门股追踪或提前发掘非主流传闻。"}
★ 277 📥 57,454