mcpField
L
llm-config

llm-config

github.com/ruvnetDevelopment80.0k741 installsUpdated today
Repository

About llm-config

A RuVLLM configuration skill for setting up local inference, task-specific MicroLoRA adapters, and SONA real-time adaptation. It guides users through checking status, generating configuration, and creating or adapting each approach.

Key features

  1. 01Check model and adapter status
  2. 02Generate RuVLLM configuration
  3. 03Create and adapt MicroLoRA adapters
  4. 04Create and adapt SONA

Use cases

  1. 01Configure local LLM inference
  2. 02Fine-tune for a specialized task
  3. 03Adapt a model using feedback signals

MCP support

Uses RuVLLM MCP tools to check status, generate configuration, and create or adapt MicroLoRA and SONA.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/ruvnet/ruflo --skill llm-config

After installing, type/llm-configin Claude Code to confirm it loaded.

Details

From the project README

LLM Configuration

Configure RuVLLM for local inference and fine-tuning.

When to use

When you need to configure local LLM inference, create MicroLoRA adapters for task-specific fine-tuning, or set up SONA for real-time adaptation.

Steps

  1. Check status — call mcp__plugin_ruflo-core_ruflo__ruvllm_status to see current model and adapter state
  2. Generate config — call mcp__plugin_ruflo-core_ruflo__ruvllm_generate_config with model parameters
  3. Create MicroLoRA — call mcp__plugin_ruflo-core_ruflo__ruvllm_microlora_create for task-specific adapters
  4. Adapt MicroLoRA — call mcp__plugin_ruflo-core_ruflo__ruvllm_microlora_adapt with training data
  5. Create SONA — call mcp__plugin_ruflo-core_ruflo__ruvllm_sona_create for real-time neural adaptation
  6. Adapt SONA — call mcp__plugin_ruflo-core_ruflo__ruvllm_sona_adapt with feedback signals

MicroLoRA vs SONA

FeatureMicroLoRASONA
SpeedMinutes to train<0.05ms adaptation
ScopeTask-specific fine-tuningReal-time micro-adjustments
PersistenceSaved as adapter weightsSession-scoped
Use caseSpecialized domain tasksContinuous feedback loops

Comments

Markdown supported