Introduction

What Problem Does Cortex TMS Solve?

When working with AI coding assistants like Claude Code or GitHub Copilot, you quickly discover a challenge: these tools are brilliant at understanding code, but they struggle when your project documentation is:

• Scattered across README files, wikis, and Notion docs
• Outdated because manual updates fell behind code changes
• Unstructured in ways AI agents can’t easily parse
• Inconsistent in format and conventions

Cortex TMS solves this by scaffolding a documentation structure that AI agents are designed to read, understand, and maintain.

Core Philosophy: Truth-Driven Documentation

Traditional documentation fails because it’s manually maintained. Cortex TMS introduces a different model:

1. Single Source of Truth: Every piece of information lives in exactly one place
2. Automated Validation: The CLI verifies documentation matches reality
3. AI-Optimized Format: Files use conventions AI agents naturally understand
4. Zero-Drift Architecture: Changes to code trigger documentation updates

Think of it as infrastructure-as-code, but for your project’s knowledge.

Key Components

Cortex TMS scaffolds these core files:

CLAUDE.md

Instructions that Claude Code reads automatically when opening your project. Defines:

• Your development workflow
• Git commit conventions
• Testing requirements
• Architecture guidelines

NEXT-TASKS.md

The current objective for you and your AI assistant. Unlike sprawling TODO lists, this focuses on:

• What’s being worked on right now
• Why it matters
• Acceptance criteria

docs/adr/

Architecture Decision Records (ADRs) documenting:

• Why you chose technology X over Y
• What trade-offs were considered
• When the decision was made

docs/core/PATTERNS.md

Code conventions AI agents should follow:

• Naming conventions
• File structure patterns
• Component design guidelines

docs/core/GLOSSARY.md

Shared vocabulary preventing confusion:

• Domain-specific terms
• Acronyms and abbreviations
• Project-specific jargon

How AI Agents Use These Files

When you ask Claude Code to “add authentication,” here’s what happens:

1. Reads CLAUDE.md to understand your workflow (e.g., “create a Git branch first”)
2. Checks PATTERNS.md to follow your naming conventions
3. Reviews GLOSSARY.md to use correct terminology
4. Updates NEXT-TASKS.md to track progress
5. Creates an ADR to document the authentication approach chosen

Without Cortex TMS, Claude would make assumptions. With it, Claude follows your team’s conventions.

Who Should Use Cortex TMS?

Perfect For:

• Solo developers using Claude Code or Copilot daily
• Small teams adopting AI-assisted development
• Open source projects onboarding AI-powered contributors
• Educators teaching LLM-augmented workflows

Not Ideal For:

• Projects with no AI tooling (use traditional docs)
• Teams committed to separate documentation sites (Cortex TMS focuses on in-repo docs)
• Highly regulated industries requiring specific doc formats

What’s Next?

Ready to try it out? Head to the Quick Start for a 5-minute tutorial, or dive into Installation for detailed setup options.