Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Calls the Didit Passive Liveness API to verify that a captured image shows a physically present person. Analyzes a single JPEG, PNG, WebP, or TIFF image and returns a liveness score, face quality rating, and luminance value. Detects spoofing attempts including printed photos, screen captures, masks, and deepfakes.
Single-image passive analysis needs no user interaction, cutting friction while maintaining 99.9% liveness detection accuracy.
When to use it
- Blocking bot signups by checking selfie liveness at registration
- Gating high-risk transactions with a real-person check
- Flagging deepfake or printed-photo fraud attempts
- Confirming physical presence before KYC identity verification
- Detecting duplicate faces across sessions to prevent account sharing
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: check whether selfie.jpg is a real person
- 1Reads selfie.jpg from disk
- 2POSTs the image to the Didit passive liveness endpoint with the DIDIT_API_KEY header and an 80-point decline threshold
- 3Parses the response liveness.status field
- 4If Declined, extracts the warnings array to identify the failure reason
- 5Returns the status, liveness score, face quality, and any warning tags
Liveness status 'Approved' with score 95, face_quality 85.0, and no warnings, or 'Declined' with a LIVENESS_FACE_ATTACK warning if spoofing is detected
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.