00-getting-started.org

LLM Command Examples

Environment Setup

These steps align with the README’s Quick Start section.

1. Initial Environment

First, ensure the environment is set up:

cd .. && make init | tee .../data/init.log | head  | tr -cd '[:print:]\n'

2. Environment Check

Verify the environment configuration:

cd .. && make check-env | tr -cd '[:print:]\n'

3. Default Model

Set up the default model:

llm models default llama3.2

Basic LLM Usage

Let’s start with a simple prompt example.

llm "Write a haiku about debugging"

Model Information

View available models and configuration.

llm models list | head -n 5

Template Management

First, list available templates.

llm templates list | head

Register a basic Python template. Note: We mark this :llm nil as it’s a one-time setup.

llm --system "Write Python code" --save python-template

Session Agent Example

Here’s an example using the session agent template.

llm -t session-agent "Reviewing LLM commands" 2>&1 | head -n 15

Command History and Logs

View recent command history

llm logs -n 10

Analyze model usage

llm logs --json -n 0 | jq -r '.[]|.model' | sort | uniq -c | sort -rn

Note on Interactive Commands

The following commands require user interaction and are not suitable for automated execution:

llm chat -m llama3.2

git diff --staged | llm -t commit