Skip to content

README.adoc

Latest commit

 

History

History
190 lines (138 loc) · 6.11 KB

README.adoc

File metadata and controls

190 lines (138 loc) · 6.11 KB

Semantic Anchors for LLMs

Table of Contents

Semantic Anchors&type=date&legend=top left

🚀 Explore the Catalog

Browse the full catalog of 60+ semantic anchors with interactive visualization, search, and role-based filtering:

What are Semantic Anchors?

Semantic anchors are well-defined terms, methodologies, and frameworks that serve as reference points when communicating with Large Language Models (LLMs). They act as shared vocabulary that triggers specific, contextually rich knowledge domains within an LLM’s training data.

Example: Instead of explaining "write tests before code, use mocks, design from outside-in," you can simply say "TDD, London School" and the LLM activates its entire knowledge about that methodology.

Why Use Them?

🎯 Precision

Reduce ambiguity by referencing established bodies of knowledge
⚡ Efficiency

Convey rich context with minimal tokens
🔄 Consistency

Ensure LLMs interpret concepts as intended
🧠 Depth

Activate multiple interconnected concepts with a single term

Quality Criteria

A good semantic anchor is:

  • Precise: References a specific, established body of knowledge

  • Rich: Activates multiple interconnected concepts

  • Consistent: Different users get similar conceptual activation

  • Attributable: Can be traced to key proponents or publications

Learn more in our About page.

Contributing

We welcome proposals for new semantic anchors! Our automated workflow with GitHub Copilot makes contributing easy:

  1. Click Propose New Anchor

  2. Provide the term and optional context

  3. Copilot validates against quality criteria

  4. If accepted, Copilot enriches and creates the anchor

  5. Maintainers review and merge the PR

See the full Contributing Guide or read it on the website.

Categories

The catalog is organized into 12 MECE (Mutually Exclusive, Collectively Exhaustive) categories:

  • Communication & Presentation

  • Design Principles & Patterns

  • Development Workflow

  • Dialogue & Interaction Patterns

  • Documentation Practices

  • Meta

  • Problem-Solving Methodologies

  • Requirements Engineering

  • Software Architecture

  • Statistical Methods & Process Monitoring

  • Strategic Planning & Decision Making

  • Testing & Quality Practices

Professional Roles

Filter anchors by role:

  • Software Developer / Engineer

  • Software Architect

  • QA Engineer / Tester

  • DevOps Engineer

  • Product Owner / Product Manager

  • Business Analyst / Requirements Engineer

  • Technical Writer / Documentation Specialist

  • UX Designer / Researcher

  • Data Scientist / Statistician

  • Consultant / Coach

  • Team Lead / Engineering Manager

  • Educator / Trainer

Project Structure

Semantic-Anchors/
├── docs/
│   ├── anchors/          # Individual anchor files (60+)
│   ├── about.adoc        # About page content
│   ├── PRD.md            # Product Requirements
│   ├── specs/            # Specifications & ADRs
│   └── arc42/            # Architecture docs
├── website/              # Vite-based static site
│   ├── src/
│   │   ├── components/   # Card Grid, Modal, etc.
│   │   ├── utils/        # Search, Router, Data
│   │   └── styles/       # Tailwind CSS
│   └── public/
├── scripts/              # Metadata extraction
├── CONTRIBUTING.adoc     # Contribution guide
└── CLAUDE.md             # Project instructions for AI agents

Development

Prerequisites

  • Node.js 20+

  • npm

Local Development

cd website
npm install
npm run dev

Build

# Extract metadata from anchors
cd scripts
npm ci
node extract-metadata.js

# Build website
cd ../website
npm run build

Deployment

The website is automatically deployed to GitHub Pages on push to main:

Technology Stack

  • Build Tool: Vite 7.x

  • Visualization: Card Grid with category sections

  • Search: Full-text search with async indexing

  • Content: AsciiDoc with asciidoctor.js

  • Styling: Tailwind CSS with typography plugin

  • i18n: English/German language switching

  • Hosting: GitHub Pages

  • Automation: GitHub Actions + Copilot

Architecture Decisions

All major decisions documented with Pugh matrices:

  • ADR-001: Vite as Static Site Generator

  • ADR-002: AsciiDoc Attributes for Metadata

  • ADR-003: Apache ECharts for Treemap (deprecated)

  • ADR-004: One File per Anchor

  • ADR-005: Card Grid Visualization (current)

License

See LICENSE file.

Maintainers

This project is maintained by the LLM Coding community.

Questions or Feedback?