Skip to content

Literate Commands

Multi-step guided workflows for opencode agents with variable collection, conditional logic, and automated script execution.

Overview

Literate commands are markdown files that define interactive, multi-step workflows. Unlike regular slash commands that execute a single action, literate commands:

  • Walk users through complex processes step by step
  • Collect structured data before proceeding
  • Branch conditionally based on user input or computed values
  • Execute scripts with interpolated variables
  • Maintain state across steps

Features

Feature Description
Multi-step workflows Break complex tasks into digestible steps
Variable collection Prompt for and store structured data
Conditional routing Branch based on conditions
Script execution Run code (bash, python, node) with variables
Variable interpolation Use $variable syntax in prompts and scripts

Security Warning

⚠️ Arbitrary Code Execution

This plugin executes arbitrary code from markdown files.

  • Literate commands can run shell scripts, Python, Node.js, and other interpreters
  • Always review command files before executing them
  • Don't run literate commands from untrusted sources
  • The plugin provides no sandboxing

Quick Start

For users: See User Guide to understand when and how to create literate commands.

For agents: See Installation to install this plugin in your project.

Documentation

Page Purpose
User Guide Learn what literate commands are and how to create them
Installation Step-by-step setup instructions for agents
Skill Instructions for synthesizing your own skill
Architecture How the plugin works (for developers)

Example

---
description: Project setup wizard
literate: true
---

```yaml {config}
step: project-name
parse:
  name: string
What would you like to name your project?

step: confirm
next:
  "confirmed === true": create
  _: cancel
Ready to create $name? Type true or false.

step: success
stop: true
✓ Project $name created! ```

Part of the opencode-literate-commands project.