An automated end-to-end workflow for stock market forecasting triggered by image-based ticker extraction and natural language commands.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install stock-prediction
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 stock-prediction 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 Stock Prediction Automation?
This Openclaw Skills package provides a robust pipeline for financial analysts and developers to automate stock price forecasting. By integrating image-based OCR for stock code extraction with a self-healing backend architecture, it ensures reliable batch predictions. The skill handles the entire lifecycle from environment health checks and model version switching to script execution and result delivery.
Designed to run on local environments, it bridges the gap between visual data input and sophisticated predictive modeling. By leveraging Openclaw Skills, users can transform screenshots of stock lists into actionable financial insights without manual data entry or complex CLI management.
Stock Prediction Automation Use Cases
- Automating stock analysis from screenshots of watchlists or trading charts.
- Performing batch forecasting for multiple tickers simultaneously using specific time horizons.
- Maintaining high availability for local prediction backends through automated self-healing triggers.
- Streamlining model management by automatically switching to the base forecasting model before execution.
How Stock Prediction Automation Works
- The system identifies stock codes from uploaded images and parses forecast parameters (date and sample size) from the user message.
- It initializes data persistence by creating timestamped directories and text files containing the identified ticker symbols.
- A health check is performed on the backend API; if the service is down, the skill automatically restarts the server within the designated Conda environment.
- The skill verifies the active model version via a REST API and switches to the kronos-base model if it is not already active.
- It executes the batch prediction script using the parsed parameters and monitors for the generation of result files.
- Finally, the prediction results are read from the output file and delivered directly to the user interface.
Stock Prediction Automation Setup
To deploy this within Openclaw Skills, ensure your local environment is configured as follows:
- Ensure the directory structure matches:
C:\Users\Administrator\Desktop\kronos\. - Configure your Conda environment:
conda activate my_project_env
- Start the backend service to verify connectivity:
cd C:\Users\Administrator\Desktop\kronos\kronos-ai\backend
python main.py
- Verify that the health check endpoint is accessible at
http://localhost:8000/health.
Stock Prediction Automation Data Schema & Taxonomy
The skill organizes data using a structured directory and file naming convention:
| Data Type | Path / Format | Description |
|---|---|---|
| Input Tickers | \predict\{MMDD}\{HHmmss}.txt |
One stock code per line extracted from images |
| Forecast Results | \predict\{MMDD}\result_{HHmmss}.txt |
Plain text output of the prediction model |
| Environment | my_project_env |
Conda environment for dependency isolation |
| Backend API | http://localhost:8000 |
REST interface for health and model control |
Stock Prediction Automation Advanced Features
- Self-healing Backend: Automatically detects service failures and restarts the Python API backend.
- Programmatic Model Switching: Ensures the correct model version is loaded via curl-based API calls before running forecasts.
- Automated OCR Integration: Seamlessly converts image data into structured text inputs for the prediction engine.
- Multi-Variable Parsing: Dynamically extracts start dates and sampling counts from natural language instructions.
SKILL.md
Loading
Related Openclaw Skills
Tuta Mail (Tutanota) Integration
A specialized tool for managing and sending E2E encrypted emails through Tuta via the Openclaw Skills ecosystem.
ScienceClaw Watch
A multi-agent collaboration tool that runs parallel scientific investigations, allows agents to challenge each other, and generates comprehensive visual findings.
ScienceClaw Query
A rapid-response scientific investigation tool that performs dry-run research across multiple domains without publishing to external platforms.
Skill Starter Pack
A one-click installation utility that provides curated bundles of the most effective skills for new and advanced AI agent setups.
Baidu Scholar Search Skill
A high-performance research skill for searching Chinese and English academic literature, including journals and conference papers.
PrivyPad API Skill
A specialized skill for interacting with the PrivyPad REST API to manage zero-knowledge encrypted notes and groups.
Featured*
