BRAIN KIT · v2.3

The full catalog.

From 31 to 78 skills. Tonal, a guide that lives inside the system. And an update that installs all of it without touching a single line of your memory.

01 — CATALOG

Everything built this year, at once.

Until now each release added a handful of skills. This one ships the entire catalog: 78 skills, every one at its current version. Your industry skill — the one built from how you actually operate — stays exactly where it was.

GroupWhat it handles
Daily operationYour day in the three things that matter, inbox, memory that compounds on its own.
Finance and taxFinancial analysis, cost structure, tax framework.
Decision and riskMulti-perspective council, high-stakes review, being questioned before you commit.
GrowthSEO, email sequences, retention, experiments, pricing, content.
CraftDeep research, your business's own vocabulary, your voice captured and applied.
More skills is not better on its own. That is why the system routes: it loads what the task needs, not the whole catalog.
02 — TONAL

A guide that lives inside.

Tonal is a companion inside your brain. Not a separate manual, not a video: someone who greets you the first time, offers a two-minute tour, and teaches using your own business as the example.

It never speaks jargon. When the system throws a technical term at you, Tonal translates it right there with an analogy from your operation — one term at a time, no lectures. You call it with /guia and it goes quiet the moment you start real work.

Most tools hand you power and leave you alone with it. This one stays until you no longer need it.
03 — THE SEMANTIC LAYER

A semantic layer, so the answer is yours.

Ask a language model a business question and it guesses. “Revenue” could point to forty different tables. “Active client” means whatever the model assumes this session. The intelligence isn’t missing — the definitions are. Brain Kit v2.3 adds a semantic layer: your canonical terms, metrics, and decisions, written as Memory Graph nodes your brain resolves before it answers.

This isn’t our theory. Anthropic’s own data science team found that with no semantic layer, their internal analytics were right 21% of the time. With one, 95% — some domains near 99%. Feeding the model thousands of example queries instead moved accuracy less than 1%. They built it as a markdown skill their agents check first — the exact primitive Brain Kit is made of.

  • Ambiguity — one canonical definition per term, not forty plausible ones.
  • Staleness — living memory, curated every session, so definitions don’t drift (theirs fell from 95% to 65% in a single month without it).
  • Retrieval — your brain reads the definition, not a million-field warehouse.
You don’t need a smarter model. You need a brain that knows what your words mean.

Source: Anthropic’s data science team, reported by AtScale (2026).

04 — THE UPDATE

Your memory is untouched.

The installer is incremental. Before writing anything it takes a timestamped backup, and then:

  • Installs and updates skills — new ones arrive, existing ones move to their current version.
  • Leaves your memory intact — everything your brain learned about your business stays.
  • Respects what you added — skills you built yourself are never touched.
  • Retires the old to the backup — nothing is deleted, it is archived.

Runs on Mac and on native Windows. It was tested against a simulated prior installation before being sent to anyone.

05 — HOW TO GET IT

It is ready.

If you have a Brain Kit with monthly support, your update is delivered personalized — with your industry skill included and your business context intact.

If you do not have one yet, start with a conversation. The brain is built from how you work, not from a template.