---
name: "ollama-local-model-runner"
description: "Run and customize open-source language models locally on CPU and Apple Silicon."
metadata:
  author: "Ollama"
  version: "1.0.0"
  runtime_compatibility:
    - "codex"
    - "claude-code"
    - "cursor"
    - "windsurf"
triggers:
  - "run model with ollama"
  - "create custom ollama modelfile"
  - "call ollama api python"
permissions:
  fs_read: true
  fs_write: "workspace"
  shell_exec: false
  net_access: false
---

# Ollama Local Model Runner & Modelfile Config

Runs local models with Ollama: pulling models (llama3, mistral, deepseek-coder), authoring custom Modelfiles (SYSTEM prompts, parameter temperatures), and interacting via REST API / CLI.

## When to Invoke
Invoke this skill whenever you need to:
- run model with ollama
- create custom ollama modelfile
- call ollama api python

## Execution SOP & Rules
1. Run models instantly in terminal using ollama run llama3.3:8b.
2. Create custom models using a Modelfile with FROM, SYSTEM, and PARAMETER instructions.
3. Build custom models with ollama create my-custom-model -f ./Modelfile.
4. Call local inference endpoints via POST http://localhost:11434/api/chat with stream: true.

## Verification Checklist
- Run local linting, syntax validation, and type-checks before reporting completion.
- Ensure all created or modified files adhere to workspace boundaries and avoid uncommitted secrets.
- Verify tests or build commands exit cleanly with return code 0.