semantic-anchors

Semantic Anchors for LLMs

A curated catalog of well-defined terms, methodologies, and frameworks that serve as reference points when communicating with Large Language Models (LLMs).

🌐 Interactive Website

Visit the Semantic Anchors website →

Features:

• 📊 Interactive Treemap Visualization - Explore 46+ semantic anchors
• 🎯 Role-Based Filtering - Filter by 12 professional roles
• 🔍 Real-Time Search - Find anchors by name, tags, or proponents
• 🌙 Dark/Light Theme - Comfortable viewing in any environment
• 🌍 Bilingual UI - English/German interface (content in English)

📖 What are Semantic Anchors?

Semantic anchors are well-defined terms, methodologies, or frameworks that act as shared vocabulary between humans and LLMs. They trigger specific, contextually rich knowledge domains within an LLM's training data.

Examples

• "TDD, London School" → Activates: outside-in testing, mock-heavy approach, interaction-based verification
• "Clean Architecture" → Activates: Uncle Bob's layered design, dependency rule, use cases as center
• "SOLID Principles" → Activates: S.O.L.I.D. breakdown, object-oriented design patterns

Why Use Them?

• Precision - Reduce ambiguity by referencing established knowledge
• Efficiency - Activate complex frameworks with minimal tokens
• Consistency - Ensure LLMs interpret concepts as the community intends
• Context Compression - Convey rich context concisely

🤖 Using Anchors with Local LLMs

Most frontier LLMs already recognize semantic anchors from their training data. However, if you use local or smaller models (Llama, Mistral, Phi, etc.) that may not know all anchors, you can inject explicit definitions into your project's agent instruction files. See the Agent Installation Guide for setup instructions covering Claude Code, Codex, Gemini CLI, Cursor, GitHub Copilot, and Windsurf.

📚 Content Structure

The catalog is organized as:

docs/
├── anchors/          # 46+ individual anchor files
│   ├── tdd-london-school.adoc
│   ├── clean-architecture.adoc
│   └── ...
├── categories/       # 11 MECE-compliant categories
├── roles/            # 12 professional roles
└── metadata/         # Generated metadata for website

🎯 Browse by Role

Find anchors relevant to your profession:

• Software Developer / Engineer
• Software Architect
• QA Engineer / Tester
• DevOps Engineer
• Product Owner / Manager
• Business Analyst / Requirements Engineer
• Technical Writer / Documentation Specialist
• UX Designer / Researcher
• Data Scientist / Statistician
• Consultant / Coach
• Team Lead / Engineering Manager
• Educator / Trainer

🤝 Contributing

We welcome contributions! See our Contributing Guide for:

• How to propose new semantic anchors
• Quality criteria for anchors
• Testing methodology with LLMs
• Development setup

Quick Start

1. Test your anchor with an LLM:

   What concepts do you associate with '<anchor name>'?
   

2. Verify quality criteria: Precise, Rich, Consistent, Attributable
3. Create an issue using our Propose New Anchor template

🛠️ Development

Prerequisites

• Node.js 20+
• npm

Setup

# Clone repository
git clone https://github.com/LLM-Coding/Semantic-Anchors.git
cd Semantic-Anchors

# Install website dependencies
cd website
npm install

# Run development server
npm run dev
# → http://localhost:5173/

# Run tests
npm run test

# Build for production
npm run build

Project Scripts

# Extract metadata from anchor files
cd scripts
npm install
node extract-metadata.js

# Validate anchors
npm run validate

# Build website
cd website
npm run build

📋 Documentation