computer-vision-expert

Guide YOLO26, SAM 3, and VLMs for vision pipeline design.

2|Updated Feb 6, 2026

Search Tags:#computer vision #object detection #segmentation #SAM #VLM #YOLO #spatial analysis

Ready to run

npx skills add https://github.com/bcastelino/agent-skills-kit --skill computer-vision-expert

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: computer-vision-expert
Source: https://github.com/bcastelino/agent-skills-kit/tree/main/skills/computer-vision-expert
Command: npx skills add https://github.com/bcastelino/agent-skills-kit --skill computer-vision-expert

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides expert guidance for designing, implementing, and optimizing state-of-the-art computer vision pipelines, from real-time object detection to advanced segmentation and spatial analysis.

Core Features & Use Cases

Unified Real-Time Detection (YOLO26): Optimized for low-latency inference and edge deployment.
Promptable Segmentation (SAM 3): Enables text-guided segmentation and 3D reconstruction.
Vision Language Models (VLMs): Facilitates semantic scene understanding and visual question answering.
Geometry & Reconstruction: Supports depth estimation, calibration, and visual SLAM.

Quick Start

Use the computer-vision-expert skill to implement a text-to-mask segmentation pipeline for identifying specific objects in an image.

Frequently Asked Questions about computer-vision-expert

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I implement text-to-mask segmentation for identifying specific objects in an image?▼

To implement text-to-mask segmentation, you use promptable segmentation models like SAM 3 to generate precise object masks from text prompts, enabling targeted identification of specific elements within an image.

What is the best way to optimize real-time object detection for edge deployment?▼

Optimizing real-time object detection for edge deployment requires using high-performance models like YOLO26, which are specifically designed for low-latency inference and efficient processing on edge devices.

Can I use vision language models for semantic scene understanding and visual question answering?▼

Yes, vision language models (VLMs) facilitate semantic scene understanding and visual question answering, allowing systems to interpret complex visual data and respond to queries about the visual content.

How do I integrate classical geometry with deep learning models for spatial analysis?▼

You integrate classical geometry with deep learning models by combining depth estimation, camera calibration, and visual SLAM techniques with modern vision systems to perform advanced spatial analysis and 3D reconstruction.

Does SAM 3 support 3D reconstruction alongside promptable segmentation?▼

Yes, SAM 3 supports 3D reconstruction alongside its core promptable segmentation capabilities, enabling text-guided mask generation that can be extended into three-dimensional spatial modeling.