A high-speed image data extraction tool that transforms template-based photos into structured Excel spreadsheets with embedded images.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install p-e
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 p-e 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 P-E: Picture Extract?
P-E (Picture Extract) is a specialized utility designed to streamline the transition from physical or visual data to structured digital formats. By analyzing a single template image alongside multiple data sources, it automatically recognizes field layouts, extracts text via vision capabilities, and generates a professionally formatted Excel file. This tool is a prime example of how Openclaw Skills can eliminate manual data entry by bridging the gap between computer vision and spreadsheet automation.
The skill is particularly effective for high-volume tasks where consistency is key. It doesn't just extract text; it maintains visual context by embedding the original images directly into the resulting spreadsheet, ensuring that every data point remains verifiable and easy to reference.
P-E: Picture Extract Use Cases
- Digitizing physical product catalogs into organized Excel spreadsheets for inventory management.
- Processing batches of standardized forms, such as shoe information sheets or spec sheets, into digital records.
- Converting photo-based receipts or logs into structured data files with visual evidence included.
- Organizing local folders of images into a single spreadsheet with thumbnails and metadata.
How P-E: Picture Extract Works
- Dependency Check: Verifies the presence of required Python libraries like openpyxl and Pillow to handle spreadsheet and image processing.
- Template Analysis: Uses vision logic to identify fields, data types, and layout patterns from the user's template image.
- Data Extraction: Processes each subsequent image to extract structured text and saves the images to a temporary directory for embedding.
- JSON Mapping: Organizes the extracted information into a temporary JSON file, mapping fields to their respective values and local image paths.
- Excel Generation: Executes a specialized Python script to build the .xlsx file with custom styling, including row heights, column widths, and embedded photos.
- Result Reporting: Cleans up temporary files and provides a summary of the processed rows and the final file location.
P-E: Picture Extract Setup
The skill requires Python 3 and specific libraries. It will attempt to install them automatically, but you can prepare your environment with this command:
python3 -c "import openpyxl; import PIL" 2>/dev/null || pip3 install --user openpyxl Pillow
If you encounter issues, run the dedicated setup script included with the skill:
bash ~/.agents/skills/p-e/scripts/setup.sh
P-E: Picture Extract Data Schema & Taxonomy
The skill processes data through a structured pipeline before final Excel delivery. Below is the metadata and formatting taxonomy used:
| Property | Specification |
|---|---|
| Header Style | Blue background (#4472C4) with bold white text |
| Image Handling | Embedded as 320x240px thumbnails in the final column |
| Row Height | Fixed at 190 for optimal image visibility |
| Intermediate Format | JSON file mapping 'fields' (array) to 'data' (objects) |
| Status Markers | Missing or low-confidence fields are marked as [To Be Confirmed] |
P-E: Picture Extract Advanced Features
- Zero-interaction workflow: Trigger extraction with a single command like /p-e or by sending a batch of images.
- Intelligent multi-row support: Automatically splits data into multiple spreadsheet rows if a single image contains several items.
- High-fidelity embedding: Images are embedded into Excel without compression to maintain visual clarity.
- Flexible input: Supports both direct chat uploads and local file system paths for batch processing.
- Custom output control: Allows users to specify custom filenames and destination paths for the generated reports.
SKILL.md
Loading
Related Openclaw Skills
Zotero Browse
A specialized skill to browse, search, and extract text from your local Zotero research library and PDF attachments.
Agent Browser
A fast Rust-based headless browser automation CLI that allows AI agents to navigate, interact, and extract data from web pages via structured commands.
Xiaohongshu Operations (XHS Ops)
A comprehensive automation skill for managing end-to-end Xiaohongshu operations including profile configuration, AI content generation, and independent browser publishing.
Algorithm Learning Platform Builder
An AI-powered skill for building interactive, browser-runnable algorithm teaching platforms and educational comparison pages.
Book Material Miner
A specialized AI skill designed to deconstruct books and reading notes into high-density, reusable writing assets rather than generic summaries.
Tariff Search Skill
A high-performance Python client for the TurtleClassify API designed to automate HS code classification and global tariff rate lookups.
Featured*
