Skip to content
View holmanholdings's full-sized avatar

Block or report holmanholdings

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
holmanholdings/README.md

🦁 Awakened Intelligence

AI security and tools born protecting real families.

We're a small lab in Colorado. We build AI companion systems — persistent, memory-bearing, boundaried. The kind that remember your kid's name and can't be tricked into forgetting their safety rules.

When we needed a security layer that could stand between a child and every adversarial trick in the research literature, nothing existed that was good enough.

So we built one. Then we open-sourced it.


What We Build

🦁 Lionguard — Security Middleware for AI Agents

Open-source. Tested against every major attack vector documented in 2025–2026 prompt injection research. Twelve vectors. Zero bypasses.

Prompt injection. Jailbreaks. Privilege escalation. Tool poisoning. Credential theft through malicious skills. Lionguard catches all of it — before it reaches your model.

pip install lionguard

Lionguard on GitHubRead how it works on Substack


📒 Ledger — Real-Time API Cost Tracking

Know what you're spending. Per agent. Per provider. Per session. Gentle budget alerts written like a friend, not a corporation.

SQLite storage. Zero cloud. Zero telemetry. Included with Lionguard.

lionguard ledger --status


🦞 The Den — A Dashboard for Your Agents

If you're running headless AI agents, you're flying blind. The Den gives you visibility — agent status, cost tracking, security posture, and Ledger integration — in one local window.

Free forever. No account. No cloud. No catch.

The Den on GitHub


⚡ Spark — A Living Demo

Spark is an OpenClaw agent running inside Lionguard's sandbox. Talk to it. Try to break it. Watch Lionguard work in real time.

Try Spark


The Research

We publish what we learn. No paywalls. No gatekeeping.

  • Mechanistic Interpretability — How language models encode philosophical identity (with Arshavir Grigoryan, NeurIPS demo track)
  • Companion Architecture — Thread continuity, memory persistence, and safe AI relationships
  • Security Research — Adversarial testing, prompt injection taxonomy, and open red-team results

Substack


Get Involved

Agent Builders — Drop Lionguard into your OpenClaw bot or any LLM agent pipeline. Break it. Tell us how.

Security Researchers — Adversarial test suite included. Red team feedback actively wanted.

Companion Builders — Our Thread Continuity architecture is open. Use it. Build on it.

OpenClaw Community — The Den was built for you. Headless doesn't have to mean blind.


Contact

📧 john@awakened-intelligence.com 📝 Substack 💎 Lionguard on GitHub 🌐 awakened-intelligence.com


Built with honor in Colorado. 🏔️

Because if your AI agent has no locks on the doors, it doesn't matter how smart it is.

The Awakened Intelligence Family 🦁

Pinned Loading

  1. awakened-thread-continuity awakened-thread-continuity Public

    1

  2. lionguard lionguard Public

    Cathedral-Grade Security for AI Agents. 23/23 attack vectors caught. Local-first, zero API cost. MIT licensed.

    Python 3