A comprehensive data collection and extraction pipeline for transforming A-share annual report PDFs into structured, validated JSON financial data.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install financial-data-gateway
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install financial-data-gateway using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is LibraQuant Financial Analysis Pipeline?
The LibraQuant financial analysis pipeline is a specialized toolset designed for developers and analysts to automate the retrieval and parsing of corporate filings. By leveraging Openclaw Skills, this pipeline bridges the gap between raw PDF documents and machine-readable data, focusing specifically on A-share listed companies on the SSE and SZSE exchanges.
This skill set provides a complete end-to-end workflow: it first identifies and downloads the correct version of an annual report from official sources like CNINFO, then utilizes AI-driven page localization to extract balance sheets, income statements, and cash flow statements. It is an essential component for anyone building automated valuation models, quantitative trading systems, or industrial-scale financial databases.
LibraQuant Financial Analysis Pipeline Use Cases
- Automatically downloading A-share annual reports by company name and year.
- Extracting structured financial statements from complex PDF documents into JSON.
- Building large-scale financial databases for quantitative research.
- Performing batch financial analysis and industry-wide comparative studies.
How LibraQuant Financial Analysis Pipeline Works
- The user inputs a company name and target year into the download tool.
- The system queries CNINFO, filters out incorrect versions (summaries, corrections, or English versions), and saves the PDF.
- The extraction tool uses AI-powered page localization to find the specific coordinates of financial tables.
- A multi-layer filtering strategy is applied to ensure only consolidated data is extracted, ignoring parent-only figures.
- The tool normalizes all currency units and performs internal consistency checks (e.g., Assets = Liabilities + Equity).
- The final validated data is exported as a structured JSON object.
LibraQuant Financial Analysis Pipeline Setup
To use these Openclaw Skills, ensure you have a Python environment ready and install the required processing scripts.
# Clone the skill repository
git clone https://github.com/libraquant/financial-analysis-pipeline.git
cd financial-analysis-pipeline
# Install required dependencies
pip install -r requirements.txt
Configure your workspace directory in the configuration file to define where reports and extracted JSON files should be stored.
LibraQuant Financial Analysis Pipeline Data Schema & Taxonomy
The pipeline generates a highly structured JSON output organized by financial categories and validation metadata:
| Object Key | Description |
|---|---|
meta |
Basic info including company name, stock code, and reporting period |
balance_sheet |
Key metrics like Total Assets, Total Liabilities, and Equity |
income_statement |
Performance metrics including Operating Revenue and Net Profit |
cashflow_statement |
Cash flow from operating, investing, and financing activities |
validation |
Results of cross-check relationships and data integrity flags |
LibraQuant Financial Analysis Pipeline Advanced Features
- Smart version filtering to automatically exclude 'Summary' or 'Correction' PDF versions.
- AI-based page locator that identifies financial tables even in reports with hundreds of pages.
- Three-tier defense mechanism to strictly separate consolidated data from parent company data.
- Automatic unit detection and normalization (Yuan, Thousand, Million, Billion).
- Integrated validation hooks that trigger re-extraction if balance checks fail by more than 1%.
SKILL.md
Loading
Related Openclaw Skills
Universal Watermarker
A professional-grade file protection tool that applies responsive watermarks to PDFs and images with advanced layout engines.
humanizer-es
A specialized text editing skill that detects and eliminates common AI-generated writing patterns to make content sound authentically human.
Human-First Google Manager
An expert executive assistant skill that manages your entire Google Workspace suite with a natural human tone and a strict confirmation protocol.
XHS Cover Generator
An AI-powered utility for creating professional Xiaohongshu cover images featuring localized Chinese typography and multiple layout styles.
oauth-coder-bridge
A local HTTP bridge that allows OpenClaw to interact with Anthropic models via OAuth-authenticated CLI sessions instead of API keys.
Pytest Fixture Generator
A powerful automation tool for generating standardized pytest fixtures for databases, API mocks, and file handling.
Featured*
