This release includes 2 security fixes for security teams reviewing exposed deployments.
Topics
+14 more
Summary
AI summaryInitial release of the WebDesign Knowledge Base Platform with AI-powered layout, motion detection, and quality evaluation.
Full changelog
ReftrixMCP v0.1.0 — 初回リリース / Initial Release
Webデザインナレッジベースプラットフォーム — AIによるレイアウト分析・モーション検出・品質評価
WebDesign Knowledge Base Platform — Layout Analysis, Motion Detection, and Quality Evaluation with AI
主な機能 / Highlights
- 19 MCPツール / 19 MCP Tools: レイアウト・モーション・品質・スタイル・ナラティブ・背景の分析 / layout, motion, quality, style, narrative, background analysis
- PostgreSQL 18 + pgvector 0.8: multilingual-e5-base Embeddingによる HNSWベクトル検索 / HNSW vector search with multilingual-e5-base embeddings
- ハイブリッド検索 / Hybrid Search: RRF(60% ベクトル + 40% 全文検索)/ RRF (60% vector + 40% full-text)
- フレーム解析 / Frame Analysis: 15px/frameスクロールキャプチャ + Pixelmatch差分検出 / 15px/frame scroll capture with Pixelmatch diff detection
- HTMLサニタイズ / HTML Sanitization: DOMPurify 3.3.xによるXSS対策 / DOMPurify 3.3.x for XSS prevention
- SSRF対策 / SSRF Protection: プライベートIP・メタデータサービスのブロック / Private IP and metadata service blocking
動作要件 / Requirements
- Node.js >= 20.19.0
- PostgreSQL 18.x + pgvector 0.8.x
- Docker(推奨 / recommended)
クイックスタート / Quick Start
git clone https://github.com/TKMD/ReftrixMCP.git
cd ReftrixMCP
pnpm install
docker compose up -d
pnpm prisma migrate deploy
pnpm build
ライセンス / License
AGPL-3.0-only
Security Fixes
- HTML sanitization implemented with DOMPurify 3.3.x to prevent XSS
- SSRF protection added by blocking private IP addresses and metadata services
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About TKMD/ReftrixMCP
Web design analysis MCP server with 26 tools for layout extraction, motion detection, quality scoring, and semantic search. Uses Playwright, pgvector HNSW, and Ollama Vision to turn web pages into searchable, structured design knowledge.
Related context
Beta — feedback welcome: [email protected]