tan-yong-shengCommunity·21 Agent Skills Included

ai-vision-mcp

Image and video analysis, object detection, design audits

Analyzes images and videos with Google Gemini or Vertex AI through a CLI or MCP server. Detects objects with bounding boxes, compares screenshots for visual regressions, and audits designs for accessibility and consistency. Removes manual screenshot review and speeds up design, testing, and content moderation workflows.

72|15

↗

npx skills add tan-yong-sheng/ai-vision-mcp --all -g -y

Available:

Explains the shared tool-function architecture so your AI agent modifies business logic in one place and keeps CLI and MCP modes consistent.

All Skills in This Repository (21)

Pure Emerald Level Indicators

📦 In Repo

tan-yong-sheng

design-system-governance

Assess design system governance and produce a structured health report.

visual-consistency-validation

Analyzes design tokens and UI components across breakpoints and color modes to identify visual inconsistencies.

playwright-cli-automation

Automate Playwright CLI screenshots, navigation flows, and Lighthouse audits.

wcag-compliance

Map accessibility findings to WCAG 2.1/3.0 criteria with remediation code examples.

design-audit-orchestration

Aggregate design audit findings into a prioritized executive summary with remediation roadmap.

web-quality-audit

Audit websites for performance, accessibility, SEO, and best practices.

investor-outreach

Create personalized investor outreach emails and follow-ups for fundraising.

market-research

Produce source-attributed market intelligence with citations and actionable recommendations.

core-web-vitals

Identify and optimize LCP, INP, and CLS targets for web pages.

article-writing

Generate long-form articles in a user's chosen voice and style.

rbac-permissions-builder

Define RBAC primitives with permission matrices and Express route guards.

opencli

Turn websites and Electron apps into programmable CLIs with YAML or TypeScript adapters.

Community

Advanced

Frequently Asked Questions

FAQPage Schema

How to install ai-vision-mcp?▼

Run `npx skills add tan-yong-sheng/ai-vision-mcp --all -g -y` in your terminal to install all skills globally.

What can ai-vision-mcp analyze?▼

It analyzes images and videos, detects objects with bounding boxes, compares up to four images, and audits UI designs for accessibility and visual consistency.

Which AI providers does it support?▼

It supports Google AI Studio (Gemini API key) and Vertex AI on Google Cloud, selectable via environment variables.

Does it work with Claude Code and Cursor?▼

Yes. It runs as a standard MCP server over stdio and also works as a standalone CLI in any terminal.

Can it analyze YouTube videos?▼

Yes. The analyze-video command accepts YouTube URLs, local files, and remote video links, with options for time ranges and frame sampling.