lorex 0.0.1-beta.2

Lorex

The Shared Knowledge Registry for AI Agents and People.

Stop repeating yourself to AI. Lorex turns your architecture notes, conventions, and runbooks into version-controlled "Skills" that every AI agent understand.

Lorex is under active development. Expect rapid changes, rough edges, and frequent improvements.

ℹ️ Who Needs Lorex?

• People using multiple AI agents and want one reusable skill source.
• Teams who want to share AI-ready project knowledge without rewriting it for every person or every tool.
• Developers who want project-specific skills to be easy for others to install and use.
• People who have strong AI sessions but find it hard to turn that knowledge into a reusable skill.

⚡ Why Lorex?

• Works with every agent: Add a skill once. Lorex projects a synchronized link to your all AI agent tools.
• Shared Intelligence: Build a central "Team Registry" (any Git repo) to share standards (e.g., security-rules, api-conventions) across your entire organization.
• Native AOT: Fast CLI, No runtime, no VM, no bulky dependencies.

🚀 Get Started

1. Install

Choose the method that fits your workflow:

For .NET Developers (Recommended)

Install Lorex as a global tool using the .NET 10 SDK:

dotnet tool install -g lorex

For Everyone Else (Standalone Binary)

Lorex is a high-performance Native binary. No runtime required.

1. Download the latest release for your OS (Windows, macOS, Linux) from GitHub Releases.
2. Add the binary to your PATH.

2. Initialize

cd your-project
lorex init

This detects your AI tools, installs the built-in lorex skill, and can suggest registry skills tagged for the current repo. If the connected registry already has skills this project does not have yet, Lorex points you to lorex install --recommended or lorex list, and lorex sync for keeping installed shared skills fresh.

🛠️ Key Use Cases

1. Let your AI write the "Lore" (Local Skills)

You don't need to write documentation manually. Because Lorex installs its own definition during init, you can simply tell your AI to document the project for you.

e.g Prompt:

"Create a lorex skill called auth-logic. Analyze our OAuth implementation and capture the core constraints, pitfalls, and flow so we don't forget them."

The Result: Your AI creates .lorex/skills/auth-logic/SKILL.md. Run lorex refresh, and that knowledge is now permanently available to every selected AI agent that opens this repo.

2. Share Your Wisdom (The Team Registry)

Turn a local skill into a company-wide standard in seconds. Lorex allows you to publish local knowledge to a shared Git registry.

# 1. Connect to your team's central library (a private Git repo)
lorex init https://github.com/your-org/ai-skills.git

# 2. Publish a locally created skill to the registry for others to use
# This makes it available to everyone else in the org!
lorex publish auth-logic

# 3. Teammates can now install it in their own repos
lorex install auth-logic

Update the skill in the registry once; lorex sync updates it for every developer and every repo in the company.

For team registries, Lorex can store a registry policy in .lorex-registry.json so contributors sync and install from the shared registry normally, while lorex publish uses a pull-request workflow instead of pushing directly.

When that policy needs to change, run lorex registry. Lorex updates the manifest directly for direct-write registries, or prepares a review branch when the registry itself is governed by a pull-request workflow.

🧠 The "Magic": How it Works

Lorex keeps one canonical source of truth for your knowledge:

.lorex/skills/
  auth-logic/
    SKILL.md
  api-conventions/
    SKILL.md

That folder is the only place Lorex expects you to author or review skill content. Everything else is a derived projection.

When you run lorex refresh, Lorex projects those skills into each agent's native integration surface.

For agents with native skill folders, Lorex creates directory symlinks back to .lorex/skills:

.claude/skills/auth-logic        -> .lorex/skills/auth-logic
.agents/skills/auth-logic        -> .lorex/skills/auth-logic
.github/skills/auth-logic        -> .lorex/skills/auth-logic
.cline/skills/auth-logic         -> .lorex/skills/auth-logic
.windsurf/skills/auth-logic      -> .lorex/skills/auth-logic
.opencode/skills/auth-logic      -> .lorex/skills/auth-logic

For agents that use rules or settings instead of skill folders, Lorex generates the right native files from the same source skill:

• Cursor → .cursor/rules/
• Roo → .roo/rules-code/
• Gemini → .gemini/settings.json

So the flow is:

1. Write or install a skill once in .lorex/skills/
2. Run lorex refresh
3. Lorex syncs every selected adapter to that same source using the format that agent already understands

Because the projections are derived from the canonical skill store, your agents stay in sync without duplicating the actual knowledge across multiple incompatible formats.

❓ Why Not Just RAG?

🤝 Contributing

Lorex is still a young project, and contributions are welcome if you want to help improve the tool, its integrations, or the overall developer experience.

git clone https://github.com/alirezanet/lorex
cd lorex
dotnet install.cs # Builds and installs the dev version

This repo dogfoods Lorex. Once cloned, your AI agent can read the lorex-contributing skill to learn the internal architecture and contribution workflow automatically!

Some future ideas we will likely explore:

• Shared prompts and other reusable AI assets alongside skills.
• Support for sub-agents and other structured agent building blocks.
• More AI providers and native integrations.
• Better ways to extract reusable skills from successful AI sessions.

License

MIT