Awesome Finance Skills

The alphaear-search skill provides unified search capabilities including web searches (via Jina, DDG, or Baidu) and local RAG (Retrieval-Augmented Generation) searches against a local daily_news database.

To perform web searches, use the SearchTools class via scripts/search_tools.py. To perform local RAG searches, use scripts/hybrid_search.py or call SearchTools with the engine parameter set to 'local'.

The alphaear-deepear-lite skill allows you to fetch high-frequency financial signals, including titles, summaries, confidence scores, and reasoning, directly from the DeepEar Lite platform.

To retrieve these signals, use the DeepEarLiteTools implementation via the scripts/deepear_lite.py script. The primary method is fetch_latest_signals(), which retrieves data from https://deepear.vercel.app/latest.json and returns a formatted report containing signal titles, sentiment/confidence metrics, summaries, and source links.

The alphaear-stock skill allows searching for A-Share, HK, and US stock tickers and retrieving historical price data (OHLCV) and fundamental information. It is accessed via the StockTools class using the scripts/stock_tools.py script.

Capabilities

• Search Tickers: Use search_ticker(query) to perform a fuzzy search by stock code or name (e.g., "Moutai" or "600519"). Returns a list of {code, name} objects.
• Get Price History: Use get_stock_price(ticker, start_date, end_date) to retrieve OHLCV data. Dates must be in "YYYY-MM-DD" format. Returns a pandas DataFrame.
• Get Fundamentals: Use get_stock_fundamentals(ticker) to retrieve a dictionary containing the sector, industry, market cap, PE ratio, and a summary. Supports A-Share, HK, and US stocks.

The alphaear-deepear-lite skill requires the following Python packages to function:

• requests
• loguru

No local database is required for this skill.

The recommended way to install individual skills is using the npx skills command. You can add a specific skill directly or search for available skills.

# Install a specific skill (e.g., alphaear-news)
npx skills add RKiding/Awesome-finance-skills@alphaear-news

# Or search for all the skills and then select one
npx skills find "alphaear"

To use the AlphaEar Predictor skill, ensure the following Python packages are installed:

• torch
• transformers
• sentence-transformers
• pandas
• numpy
• scikit-learn

The AlphaEar Predictor skill uses the KronosPredictorUtility to perform time-series forecasting. The workflow involves generating a base quantitative forecast and then subjectively adjusting it using a specific prompt (found in references/PROMPTS.md) to account for news sentiment.

Key Method:

• KronosPredictorUtility.get_base_forecast(df, lookback, pred_len, news_text): Returns a List[KLinePoint] based on technical data and news input.

from scripts.utils.kronos_predictor import KronosPredictorUtility
from scripts.utils.database_manager import DatabaseManager

db = DatabaseManager()
predictor = KronosPredictorUtility()

# Forecast
forecast = predictor.predict("600519", horizon="7d")
print(forecast)

When higher accuracy or reasoning is required, use an LLM to analyze the text directly using a specific prompt. After analysis, use the provided helper method to save the results to the database.

Sentiment Analysis Prompt:

请分析以下金融/新闻文本的情绪极性。
返回严格的 JSON 格式:
{"score": <float: -1.0到1.0>, "label": "<positive/negative/neutral>", "reason": "<简短理由>"}

文本: {text}

Scoring Guide:

• Positive (0.1 to 1.0): Optimistic news, profit growth, policy support, etc.
• Negative (-1.0 to -0.1): Losses, sanctions, price drops, pessimism.
• Neutral (-0.1 to 0.1): Factual reporting, sideways movement, ambiguous impact.

Helper Method to save results:

• update_single_news_sentiment(id, score, reason): Use this to save your manual LLM analysis to the database.

To use the AlphaEar Signal Tracker pattern, ensure the following dependencies are available in your environment:

• agno (Agent framework)
• sqlite3 (built-in Python module)

Note: Ensure that DatabaseManager is initialized correctly within your implementation.

Match the level of specificity in your skill to the task's variability:

This skill allows an agent to create visual representations of finance logic flows, such as investment theses or signal transmission chains, by generating Draw.io XML compatible diagrams.

Agentic Workflow

To use this skill in an agentic workflow, follow these steps:

1. Generate XML: Use the Draw.io XML Generation Prompt (found in references/PROMPTS.md) to convert a logical chain into XML format.
2. Save/Render: Call the render_drawio_to_html method from scripts/visualizer.py to convert the XML into a viewable HTML file for the user.