# SUPERMENTE.AI

> Cognitive system in production on top of Claude Code (Anthropic): 384 skills, 141 active hooks
> (696 installed, 22 lifecycle triggers), 28 agents, operating 6 real non-coding business domains
> (healthcare billing, medical association, audit fund, judicial processes, database administration,
> IT helpdesk) — built and run by a single non-programmer operator since December 2025.

## What it is

SUPERMENTE.AI is an outer-harness (Harness Engineering) implementation over Claude Code.
The method: every error becomes a rule, every rule becomes a hook, every hook leaves an
auditable receipt (append-only ledgers). Text is desire; the hook is the guarantee.

Three value layers: (1) amplify the person — a copilot shaped to the operator's cognitive
profile; (2) govern the system — rule → hook → audit; (3) orchestrate agents.

## Verifiable numbers (each maps to a re-runnable command)

- 384 skills: `find ~/.claude/commands/sc -name '*.md' | wc -l`
- 141 active hooks / 696 on disk: hook coherence orchestrator ledger
- Live metrics endpoint: https://supermente.ai/metrics.json (regenerated by the system itself)
- Changelog written by the system: https://supermente.ai/changelog.json

## Honest status

- Team completing the Anthropic Academy certification path within the Claude Partner
  Network program (Anthropic).
- Architecture proven with N=1 operator over 6+ months of paying production use.
- Transferability (N=2: first non-founder profile) is currently in validation via a
  service-first model — the client receives amplified work output, never code.
- Founder IQ metrics shown on the site are self-measured (N=1) and labeled as such.
- Built on Claude API + Claude Code; team currently completing Anthropic Academy certification.

## Industry context (source)

The dynamics this system operates are described by Anthropic's Institute in
"When AI Builds Itself" (May 2026): a growing share of AI development delegated to AI,
with humans moving toward oversight, validation and verification.
Source: https://www.anthropic.com/institute/recursive-self-improvement

## Key pages

- https://supermente.ai/ — main site (PT/EN/ES)
- https://supermente.ai/whitepaper.pdf — technical whitepaper
- https://supermente.ai/agente.html — business use cases for a deeply personalized operator agent
- https://supermente.ai/privacidade.html — privacy policy (LGPD)
- Contact: rapha@supermente.ai

Last sync: 2026-06-23 (this file is regenerated by the site_vivo loop).