neural-training

Original：🇺🇸 English

Translated

Neural pattern training with SONA (Self-Optimizing Neural Architecture), MoE (Mixture of Experts), and EWC++ for knowledge consolidation. Use when: pattern learning, model optimization, knowledge transfer, adaptive routing. Skip when: simple tasks, no learning required, one-off operations.

1installs

Sourceruvnet/claude-flow

Added on2026-02-08

NPX Install

npx skill4agent add ruvnet/claude-flow neural-training

SKILL.md Content

View Translation Comparison →

Neural Training Skill

Purpose

Train and optimize neural patterns using SONA, MoE, and EWC++ systems.

When to Trigger

Training new patterns
Optimizing agent routing
Knowledge consolidation
Pattern recognition tasks

Intelligence Pipeline

RETRIEVE — Fetch relevant patterns via HNSW (150x-12,500x faster)
JUDGE — Evaluate with verdicts (success$failure)
DISTILL — Extract key learnings via LoRA
CONSOLIDATE — Prevent catastrophic forgetting via EWC++

Components

Component	Purpose	Performance
SONA	Self-optimizing adaptation	<0.05ms
MoE	Expert routing	8 experts
HNSW	Pattern search	150x-12,500x
EWC++	Prevent forgetting	Continuous
Flash Attention	Speed	2.49x-7.47x

Commands

Train Patterns

bash

npx claude-flow neural train --model-type moe --epochs 10

Check Status

bash

npx claude-flow neural status

View Patterns

bash

npx claude-flow neural patterns --type all

Predict

bash

npx claude-flow neural predict --input "task description"

Optimize

bash

npx claude-flow neural optimize --target latency

Best Practices

Use pretrain hook for batch learning
Store successful patterns after completion
Consolidate regularly to prevent forgetting
Route based on task complexity

neural-training

NPX Install

Tags

SKILL.md Content

Neural Training Skill

Purpose

When to Trigger

Intelligence Pipeline

Components

Commands

Train Patterns

Check Status

View Patterns

Predict

Optimize

Best Practices