Hurricane Learner in Plain English v1
This is the stored snapshot for the approved document version. The diff below shows what changed from the previous version.
Document snapshot
Hurricane Learner in Plain English
Summary
Learner teaches an AI a topic from your own sources. You hand it the articles, videos, and PDFs you trust; it reads them, remembers what matters, and writes the AI a skill it can use — plus it keeps a searchable memory of everything you fed it, so you can ask questions and get answers grounded in your material.
How you'd use it
You pick a topic and give Learner some sources — a few article links, a YouTube video, a PDF report, or just text you paste in. Learner goes through each one, pulls out the parts worth keeping (the rules, the steps, the definitions, the warnings), and files it all away in a private library on your own machine. That library sticks around: add more sources next week and it just grows.
Once it has enough, you can do two things with it:
- Ask it questions. "What does my library say about X?" — and it answers using only what you gave it, and tells you which sources it leaned on.
- Turn it into a skill. With one command, Learner boils everything it learned about the topic down into a single tidy playbook — a cheat-sheet of the key ideas and a long list of "if this, then do that" rules. That playbook is exactly the shape Hurricane Sentinel expects, so you can hand it to one of your AI agents and it just works. The reading you did once becomes something the agent knows forever.
The idea
Most AI assistants only know what they were trained on, plus whatever you paste in each time. Learner lets you build up a lasting body of knowledge on the topics you care about, from the sources you trust — and then package that knowledge so an AI agent carries it into every job, instead of starting from scratch every time.
Where it's at
Learner is being built. It's an internal tool right now, run from a command line on Linux, and it needs a local AI backend (like Ollama) to do the reading and writing. The core is working today: feed it sources, keep a searchable library, ask it questions, and build a ready-to-use skill from what it learned. Wiring it directly into Sentinel and driving it from a browser are next.