SkillsAdd

davidcastagnetoa/skills

silent_face_anti_spoofing

Modelo de liveness pasivo NUAA para detectar ataques de impresión, pantalla y spoofing sin interacción del usuario

First seen Mar 6, 2026

Installation

$ npx skills add davidcastagnetoa/skills --skill silent_face_anti_spoofing
View on skills.sh ↗ GitHub repository ↗

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Device Fingerprint Spoofing Similar name
mukul975/anthropic-cybersecurity-skills

Performs OAuth 2.0 scope minimization review to identify over-permissioned third-party applicat…

113 installs

Also in this package

Other skills from davidcastagnetoa/skills · top by installs.

npx skills add davidcastagnetoa/skills
traefik 116 prisma-nestjs-patterns 58 easyocr 50 exif_metadata_analyzer 30 c4_model_structurizr 28 configurable_thresholds 25 alembic 23 insightface_arcface 19 asyncpg 19 github_actions_cicd 16

Browse all from davidcastagnetoa/skills

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 1
Default branch main
Open issues 0
Status Active

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 1,463 B
  • docs SUMMARY.md 149 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 5 installs

SKILL.md

silentfaceanti_spoofing

Silent-Face-Anti-Spoofing (NUAA) es el modelo principal de liveness pasivo. Analiza un único frame para determinar si el rostro es real o un ataque (foto impresa, pantalla, máscara).

When to use

Ejecutar sobre cada frame de selfie antes del challenge activo. Es la primera línea de defensa contra spoofing.

Instructions

  1. Clonar el repositorio: git clone https://github.com/minivision-ai/Silent-Face-Anti-Spoofing.git
  2. Instalar dependencias: pip install torch torchvision opencv-python.
  3. Descargar los pesos preentrenados del repositorio oficial (modelos 2.780x80 y 40080x80).
  4. Cargar ambos modelos en memoria al arrancar el worker (model warm-up).
  5. Preprocesar el frame: recortar región facial, redimensionar a 80x80, normalizar.
  6. Ejecutar inferencia en ambos modelos y promediar scores.
  7. Umbral de liveness: score > 0.6 = real; score ≤ 0.6 = spoof.
  8. Exportar a ONNX para despliegue en Triton: torch.onnx.export(model, dummyinput, 'silentfas.onnx').

Notes