i-have-adhd — ADHD-Friendly Output Skill for Coding Agents

Introduction

i-have-adhd is a skill for AI coding agents that restructures their output to be concise, action-oriented, and scannable. Long-winded explanations bury the actual answer, which is especially frustrating for developers with ADHD or anyone who wants to get straight to the point. This skill instructs agents to lead with the action, use numbered steps, and always end with a concrete next step.

What i-have-adhd Does

• Forces coding agent output into a structured, action-first format
• Eliminates verbose preambles and unnecessary context before the answer
• Adds numbered steps so responses are scannable at a glance
• Ensures every response ends with a clear, concrete next step
• Works with Claude Code, Cursor, Codex, and other agent frameworks

Architecture Overview

The skill is a prompt-based instruction set that integrates with AI coding agent frameworks. It does not modify the underlying model but shapes output formatting through carefully crafted system-level instructions. The skill file is loaded into the agent's context at the start of each session, guiding how the agent structures its responses throughout the conversation.

Self-Hosting & Configuration

• Clone the repository or copy the skill file into your agent's skill directory
• For Claude Code, place the skill in .claude/skills/ within your project
• For Cursor, add it to your rules or system prompt configuration
• No external dependencies or services required
• Customize the formatting rules by editing the skill prompt directly

Key Features

• Action-first formatting that puts the answer before the explanation
• Numbered step lists for easy scanning and task tracking
• Concrete next-step suggestions at the end of every response
• Framework-agnostic design that works across major AI coding tools
• Lightweight prompt-only implementation with zero runtime overhead

Comparison with Similar Tools

• Default agent output — verbose, buries the answer in context and caveats
• Custom system prompts — requires manual prompt engineering per project
• Hallmark — focuses on anti-AI-slop design; i-have-adhd focuses on output structure
• Claude CLAUDE.md — project-level config; i-have-adhd is a portable skill
• Cursor rules — editor-specific; i-have-adhd works across multiple agents

FAQ

Q: Does this change the quality of the agent's code output? A: No. It only changes how the response is structured and presented, not the underlying reasoning or code generation.

Q: Which AI coding agents are supported? A: Any agent that supports custom skills or system prompts, including Claude Code, Cursor, GitHub Copilot, and OpenAI Codex CLI.

Q: Can I customize the formatting rules? A: Yes. The skill is a plain text prompt file that you can edit to match your preferred output style.

Q: Is this only useful for people with ADHD? A: No. Anyone who prefers concise, action-oriented responses benefits from this skill. The name reflects its origin but the utility is universal.

Sources

• https://github.com/ayghri/i-have-adhd