A high-precision liveness detection skill that identifies whether a face in a static image is a real person or a fraudulent attempt like a screen capture or 3D mask.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install tencentcloud-faceid-detectlivefaceaccurate
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 tencentcloud-faceid-detectlivefaceaccurate 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 Tencent Cloud FaceID DetectLiveFaceAccurate?
The Tencent Cloud FaceID DetectLiveFaceAccurate skill is a professional-grade security tool designed for high-precision static liveness detection. By integrating this into Openclaw Skills, developers can effectively defend against sophisticated presentation attacks, including high-definition screen re-shoots, printed paper cutouts, and 3D masks. It is significantly more robust than standard static detection methods, making it ideal for high-security mobile and PC applications.
This skill supports flexible input methods, accepting both local file paths (with automatic Base64 conversion) and direct image URLs. It provides a granular liveness score and a clear boolean result, allowing for customizable security thresholds based on specific business needs.
Tencent Cloud FaceID DetectLiveFaceAccurate Use Cases
- Verifying if a user-uploaded profile photo is a live person rather than a photo of a screen.
- Preventing identity fraud in financial or KYC (Know Your Customer) workflows.
- Detecting high-tech spoofing attempts such as 3D masks or high-resolution printed photos.
- Implementing secure remote authentication for mobile and web platforms.
How Tencent Cloud FaceID DetectLiveFaceAccurate Works
- The user provides a static image via a local file path, a Base64 string, or a public URL.
- The skill processes the image, ensuring it meets resolution and size specifications (e.g., face size > 100x100 pixels).
- It calls the Tencent Cloud DetectLiveFaceAccurate API using authenticated credentials.
- The AI model analyzes the image for depth, texture, and reflection patterns typical of re-shot or synthetic media.
- The skill returns a JSON response containing a liveness score (0-100) and a recommendation based on a standard threshold of 40.
Tencent Cloud FaceID DetectLiveFaceAccurate Setup
To use this skill within Openclaw Skills, ensure you have Python 3.6+ and the required SDK installed:
pip install tencentcloud-sdk-python
Set the following environment variables with your Tencent Cloud credentials:
export TENCENTCLOUD_SECRET_ID="your_secret_id"
export TENCENTCLOUD_SECRET_KEY="your_secret_key"
Tencent Cloud FaceID DetectLiveFaceAccurate Data Schema & Taxonomy
The skill accepts specific image formats and returns detailed metadata as follows:
Input Specifications
| Feature | Requirement |
|---|---|
| Format | PNG, JPG, JPEG, BMP (GIF not supported) |
| Max Size | 5MB after Base64 encoding |
| Face Size | Minimum 100x100 pixels |
| Aspect Ratio | Recommended 3:4 |
Output Metadata
| Key | Description |
|---|---|
| Score | Liveness score from 0 to 100 |
| IsLive | Boolean indicating if the score meets the threshold |
| ScoreDesc | Textual description of the result |
| RequestId | Unique identifier for the API call |
Tencent Cloud FaceID DetectLiveFaceAccurate Advanced Features
- Automatic Base64 Encoding: Simplifies development by automatically converting local image files into the required string format.
- Customizable Thresholds: While the default threshold is 40, users can adjust sensitivity (5, 10, 70, 90) based on risk tolerance.
- Multi-Version Support: Compatible with algorithm model version 3.0 for the latest in facial recognition technology.
- Region Specification: Allows users to define specific Tencent Cloud regions for optimized latency and compliance.
SKILL.md
Loading
Related Openclaw Skills
Draw Animal Skill
A Python-powered utility for generating descriptive text for animal imagery based on user-defined parameters.
Insomnia Audio Sleep Aid
A dedicated AI skill providing empathetic guidance and curated audio resources for users experiencing insomnia and sleep difficulties.
GitHub Development Standard
A rigorous 9-step development framework and validation system designed to ensure AI agents produce high-quality, verified code within GitHub repositories.
Reddit Market Insights
An AI-powered semantic search tool for uncovering ecommerce opportunities, buyer pain points, and trending products directly from Reddit discussions.
Chemical Storage Sorter
A specialized tool for classifying laboratory chemicals and organizing them into safe storage groups based on compatibility and hazard regulations.
LobsterDomains
A high-performance AI agent skill for searching and registering domain names using cryptocurrency payments across multiple blockchain networks.
Featured*
