Skill
NEXO Brain
Cognitive memory system for AI agents — Atkinson-Shiffrin memory model, semantic RAG, trust scoring, and metacognitive error prevention. Gives your agent per...
When to use NEXO Brain
Choose if
You want your agent to have persistent, locally-stored memory across sessions with semantic search, an Atkinson-Shiffrin-style decay model, pre-edit guard checks against past mistakes, and session diaries. Pick this over a vector-DB-only RAG when you also need episodic logs, learning rules, and a metacognitive trust signal — and want everything on-disk with no cloud calls.
Avoid if
You need cloud-synced memory across multiple machines or users (NEXO is local-only per the README), or a Windows host (latest version lists Darwin and Linux only). Also avoid when your agent only needs simple vector retrieval — the cognitive-engine scope is broader than a plain RAG store and adds setup complexity.
Risk Flags
- MEDIUM lock-in Memory is stored in two SQLite databases under ~/.nexo/; migrating to another memory system requires custom export logic. README mentions a backup/retention feature but no documented portable export format.
- LOW scope README scopes the system to single-host local memory (Darwin/Linux). No cloud sync or multi-device replication is claimed.
Cost
Type: Free · Free tier: README states everything stays local — two SQLite databases in ~/.nexo/, no telemetry, no cloud APIs.
Distribution
- ClawHub
nexo-brain- Repository
- https://github.com/wazionapps/nexo