# OpenCLI Agent Pack: LLM

Use this when an AI agent needs to work with `llm`.

## What this CLI is for
Prompting, local plugins, and structured llm outputs from the terminal. Built by Simon Willison. Supports structured output — good for scripts and agents.

Best for: prompting, local plugins, and structured llm outputs from the terminal.

## Agent readiness
Great for agents (75/100)
- Structured output is available for parsing.
- Supports non-interactive/scripted use.
- Works well in CI or repeatable automation.

## Install
```sh
pipx install llm
```

## Verify before real work
```sh
llm --help
```
Expected signal: llm responds locally; authenticate before real work.

## Safe starting commands
```sh
llm --help
```

```sh
llm 'Explain this command: rg TODO src'
```

## Guardrails for agents
- Verify identity/account before running task commands.
- Network access is required; avoid leaking secrets in logs.
- Start read-only, then ask before mutations.

## Suggested agent instruction
You may use LLM (`llm`) for prompting, local plugins, and structured llm outputs from the terminal.. First install it if missing, then run the verify command. Start with read-only or inspection commands. Summarize what you found before changing anything. Ask for confirmation before commands that mutate remote state, spend money, deploy, delete data, merge code, or expose secrets.

Source: OpenCLI