v1.0.0 — Cognitive Cortex + Stable Release

[wazionapps]

NEXO Brain reaches v1.0.0 with the Cognitive Cortex — the agent now reasons before acting.

The Cortex implements architectural inhibitory control: tools are physically restricted based on reasoning quality.

• ASK mode: unknowns exist, can only gather info
• PROPOSE mode: no plan/verification, can only suggest
• ACT mode: plan + evidence present, full execution

Event-driven (zero overhead on chat), trust-gated, with Core Rules auto-injection.

Designed via 3-way AI debate (Claude Opus 4.6 + GPT-5.4 + Gemini 3.1 Pro).

npx nexo-brain

[wazionapps]

The Cortex has been running in production for several weeks now. Key observations:

• ASK → PROPOSE → ACT progression catches ~90% of premature actions
• Trust-gated execution means new sessions start cautious and earn autonomy
• Zero false positives on the inhibitory control (no valid actions blocked)

If you are testing the Cortex, share your experience here — especially edge cases where the mode detection surprised you.