How to Use the /i-have-adhd Command: A Complete Guide

The /i-have-adhd command activates a specialized skill that restructures AI responses with ten specific formatting rules designed for ADHD readers, including leading with next actions, numbering steps, and providing time estimates.

This guide explains how to enable and use the ADHD formatting mode in the ayghri/i-have-adhd repository. The command transforms standard assistant output into concise, action-oriented responses that improve focus and reduce cognitive load.

What the /i-have-adhd Command Does

When invoked, the command loads the i-have-adhd skill, which applies a strict formatting protocol defined in skills/i-have-adhd/SKILL.md. The skill enforces ten concise rules: leading every response with the next immediate action, numbering all procedural steps, providing time estimates for tasks, and breaking complex information into scannable chunks. This reshaping happens at the system prompt level, ensuring every subsequent response follows ADHD-friendly formatting without requiring repeated instructions.

Installation and Setup

The skill is distributed as a plugin for multiple AI runtimes including Claude, Codex, Pi, OMP, Qwen, and Gemini. Each platform uses the same core SKILL.md file, guaranteeing identical behavior across environments.

OpenCode Plugin Registration

For OpenCode users, the plugin file .opencode/plugins/i-have-adhd.mjs handles command registration. This module registers the slash command and manages the runtime configuration by adding the skill's path to config.skills.paths. To install, add the plugin entry to your opencode.json configuration:

{
  "plugin": ["./.opencode/plugins/i-have-adhd.mjs"]
}

Restart OpenCode after modifying the configuration. The plugin also references .opencode/command/i-have-adhd.md for command description metadata used in the UI.

Activating ADHD Mode

You can activate the formatting mode either for a single session or persistently across all conversations.

Session-Only Activation

Type the slash command in any active chat session:

/i-have-adhd

This immediately loads the skill rules from SKILL.md into the current context. The assistant will confirm activation and begin formatting all subsequent responses according to the ten-rule protocol.

Always-On Mode

For persistent activation without typing the command repeatedly, create a flag file in your configuration directory. The plugin checks for this file on every turn and automatically injects the full rule set into the system prompt.

For OpenCode:

touch ~/.config/opencode/.i-have-adhd-always

For Claude:

touch ~/.claude/.i-have-adhd-always

The hooks/always-on.mjs file contains the system-prompt transformer logic that enables this behavior. This hook is shared across runtimes, ensuring consistent always-on functionality regardless of which AI client you use.

Disabling the Mode

To return to standard assistant formatting, say either "stop adhd mode" or "normal mode". The assistant confirms the change in a single line and immediately reverts to default response styles. To disable always-on mode while keeping the command available for manual use, delete the flag file:

rm ~/.config/opencode/.i-have-adhd-always

Core Rules and Source Files

The canonical behavior is defined in skills/i-have-adhd/SKILL.md, which serves as the single source of truth for all runtimes. This file contains the ten specific formatting instructions that the plugin enforces. The system never deviates from these rules unless explicitly deactivated.

Key source files include:

  • skills/i-have-adhd/SKILL.md – Contains the ten rule definitions governing output formatting
  • .opencode/plugins/i-have-adhd.mjs – OpenCode plugin handling command registration and skill path injection
  • hooks/always-on.mjs – System-prompt transformer enabling automatic rule injection for always-on mode
  • INSTALL.md – Platform-specific installation instructions and flag-toggle commands

Usage Examples

Install and configure the plugin for immediate use:


# 1. Install the plugin (OpenCode example)

# Add to opencode.json and restart OpenCode

# 2. Activate ADHD mode for current session only

/i-have-adhd

# 3. Enable persistent always-on mode

touch ~/.config/opencode/.i-have-adhd-always

# 4. Disable always-on mode (command remains available)

rm ~/.config/opencode/.i-have-adhd-always

Summary

  • The /i-have-adhd command activates a skill that reformats all assistant responses according to ten specific ADHD-friendly rules.
  • Installation requires adding the platform-specific plugin (such as .opencode/plugins/i-have-adhd.mjs) to your runtime configuration.
  • Use the slash command for temporary activation, or create the .i-have-adhd-always flag file for persistent mode.
  • Deactivation occurs automatically when you type "stop adhd mode" or "normal mode".
  • All behavior is governed by skills/i-have-adhd/SKILL.md, ensuring consistent formatting across Claude, OpenCode, and other supported runtimes.

Frequently Asked Questions

Where are the formatting rules defined?

The formatting rules live in skills/i-have-adhd/SKILL.md in the repository root. This file is the single source of truth that all runtimes reference, ensuring identical ADHD-friendly formatting whether you use Claude, OpenCode, or other supported platforms.

How do I keep ADHD mode enabled permanently?

Create a flag file in your runtime's configuration directory. For OpenCode, run touch ~/.config/opencode/.i-have-adhd-always. For Claude, use ~/.claude/.i-have-adhd-always. The hooks/always-on.mjs transformer automatically injects the skill rules on every conversation turn when this file exists.

Will the command work on all AI platforms?

Yes. The repository distributes platform-specific plugins for Claude, Codex, Pi, OMP, Qwen, Gemini, and others. While installation methods vary (documented in INSTALL.md), each runtime loads the same SKILL.md rules, guaranteeing consistent behavior across all supported clients.

How do I turn off ADHD formatting?

Say "stop adhd mode" or "normal mode" in the chat. The assistant confirms the change immediately and reverts to standard response formatting. To disable always-on mode permanently (while keeping the manual command functional), remove the flag file from your configuration directory.

Have a question about this repo?

These articles cover the highlights, but your codebase questions are specific. Give your agent direct access to the source. Share this with your agent to get started:

Share the following with your agent to get started:
curl -s "https://instagit.com/install.md"

Works with
Claude Codex Cursor VS Code OpenClaw Any MCP Client

Maintain an open-source project? Get it listed too →