loops-cli

---
name: loops-cli
description: >
license: MIT
---

# Loops CLI Skill

This skill helps with Loops terminal workflows. Use it for installation, auth and configuration, command selection, and shell-first operational tasks.

## When To Use

Use this skill when the user needs to:

- install, update, or troubleshoot the Loops CLI
- authenticate with Loops from the terminal
- manage stored team keys or switch between them
- run one-off contact, list, event, or transactional-email commands
- create draft campaigns, update email-message content, and upload images from the shell
- list/get themes and reusable components for LMX
- inspect CLI output in text or JSON locally

This skill is for command-line usage, not application integrations or email-strategy review.

## Working Style

When this skill is active:

1. Prefer `loops agent-context` for exact flags and the latest command shape.
2. Prefer the CLI for shell workflows, one-off operational tasks, credential validation, and quick troubleshooting.
3. Use `--output json` when the result needs to feed another tool or script.
4. Use named stored keys plus `--team` when the user works across multiple Loops teams.
5. Avoid printing secrets. Prefer keyring-backed auth or environment variables over hardcoded API keys.
6. If the task becomes application-code integration or exact HTTP payload design, use the separate `loops-api` skill.

Official references:

- CLI docs: `https://loops.so/docs/cli`
- CLI repo: `https://github.com/loops-so/cli`
- CLI README: `https://github.com/loops-so/cli/blob/main/README.md`

## Category Routing

- Installation, auth flows, config resolution, global flags, and common Loops CLI workflows:
  Read `references/cli.md`
- Campaigns, email messages, themes, components, revision handling, LMX file flags, and uploads:
  Read `references/cli.md`. For LMX markup itself, also use the `loops-lmx` skill.

If the task becomes application-code integration or exact HTTP payload design beyond the CLI, use the `loops-api` skill.

## Output Checklist

Aim to leave the user with:

- the right command or install path for the task
- any auth or team-selection caveats that affect behavior
- safe handling of credentials
- the next validation step, such as `loops --help`, `loops auth status`, `loops api-key`, or `loops agent-context`