Neural Training
- 1k installs
- 67k repo stars
- Updated August 4, 2026
- ruvnet/ruflo
neural-training is an agent skill that continuously trains, optimizes, and consolidates neural patterns inside AI agents for developers who need better routing, pattern retention, and adaptive knowledge transfer.
About
neural-training is a ruvnet/ruflo agent skill that trains and optimizes neural patterns using SONA (Self-Optimizing Neural Architecture), MoE (Mixture of Experts), and EWC++ for knowledge consolidation. The intelligence pipeline retrieves patterns via HNSW indexing reported as 150x–12,500x faster retrieval, judges quality, and consolidates learned routing behavior for agent systems. Developers reach for neural-training when building agents that must learn recurring patterns, optimize routing decisions, or retain knowledge across sessions without full retraining. Skip it for one-off tasks with no learning requirement.
- SONA self-optimizing neural architecture for real-time adaptation
- Mixture-of-Experts (MoE) routing with 8 experts
- HNSW vector search delivering 150x–12,500x faster pattern retrieval
- EWC++ consolidation to prevent catastrophic forgetting
- Full intelligence pipeline: RETRIEVE → JUDGE → DISTILL → CONSOLIDATE
Neural Training by the numbers
- 1,014 all-time installs (skills.sh)
- +3 installs in the week ending Aug 5, 2026 (Skillselion tracking)
- Ranked #1,036 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Security screen: LOW risk (skills.sh audit)
- Data as of Aug 5, 2026 (Skillselion catalog sync)
npx skills add https://github.com/ruvnet/ruflo --skill neural-trainingAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 1k |
|---|---|
| repo stars | ★ 67k |
| Security audit | 3 / 3 scanners passed |
| Last updated | August 4, 2026 |
| Repository | ruvnet/ruflo ↗ |
How do you train agent routing patterns with consolidation?
Continuously train, optimize, and consolidate neural patterns inside their AI agents for better routing and long-term knowledge retention.
Who is it for?
Developers building adaptive AI agents that need persistent pattern learning, routing optimization, and knowledge transfer.
Skip if: Developers handling one-off prompts or workflows with no recurring patterns to learn or retain.
When should I use this skill?
A developer asks to train agent patterns, optimize routing, consolidate knowledge, or improve long-term agent memory behavior.
What you get
Trained neural patterns, optimized agent routing rules, and consolidated long-term knowledge representations.
- trained neural patterns
- optimized routing configuration
By the numbers
- HNSW pattern retrieval reported as 150x–12,500x faster
- Uses three training systems: SONA, MoE, and EWC++
Files
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
1. RETRIEVE — Fetch relevant patterns via HNSW (150x-12,500x faster) 2. JUDGE — Evaluate with verdicts (success$failure) 3. DISTILL — Extract key learnings via LoRA 4. 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
npx claude-flow neural train --model-type moe --epochs 10Check Status
npx claude-flow neural statusView Patterns
npx claude-flow neural patterns --type allPredict
npx claude-flow neural predict --input "task description"Optimize
npx claude-flow neural optimize --target latencyBest Practices
1. Use pretrain hook for batch learning 2. Store successful patterns after completion 3. Consolidate regularly to prevent forgetting 4. Route based on task complexity
Related skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10How it compares
Pick neural-training when agents must learn and retain routing patterns over time, not when you only need static prompt templates.
FAQ
What systems does neural-training use?
neural-training applies SONA (Self-Optimizing Neural Architecture), MoE (Mixture of Experts), and EWC++ for knowledge consolidation inside ruflo agent workflows. Pattern retrieval runs through an HNSW pipeline described as 150x–12,500x faster.
When should I skip neural-training?
neural-training is unnecessary for simple one-off agent tasks with no learning requirement. The skill readme explicitly recommends skipping it when pattern retention, routing optimization, or knowledge transfer are not goals.
What problems does neural-training solve?
neural-training addresses training new patterns, optimizing agent routing, knowledge consolidation, and pattern recognition inside multi-session agent systems. Outputs include improved routing decisions and retained learned behaviors.
Is Neural Training safe to install?
skills.sh reports 3 of 3 security scanners passed. Review the Security Audits panel on this page before installing in production.