How to Customize the 10 ADHD-Friendly Rules by Forking the i-have-adhd Skill

To customize the 10 ADHD-friendly response rules, fork the ayghri/i-have-adhd repository, edit the canonical skill file at skills/i-have-adhd/SKILL.md, synchronize the Cursor-compatible mirror at .cursor/skills/i-have-adhd/SKILL.md, and republish your fork to the marketplace so runtimes like Claude Code or Codex load your modified rules instead of the upstream defaults.

The ayghri/i-have-adhd plugin governs LLM output using a skill-based architecture that enforces formatting constraints at runtime. Because the rule set resides in a standalone markdown file rather than compiled code, you can fully customize the 10 default guidelines—or add your own—by forking the repository and modifying the skill definition.

Locate the Canonical Rule Definition

The single source-of-truth for response behavior is skills/i-have-adhd/SKILL.md in the repository root. This file contains the 10 ADHD-friendly rules that dictate how the AI structures output, covering principles such as information chunking, bold text highlighting, and avoidance of dense text walls. When the plugin executes inside Claude Code, Codex, Pi, OMP, or OpenCode, the runtime reads this specific file from the marketplace distribution to determine formatting constraints.

Fork and Modify the Skill

Step 1: Fork the Repository

Create your own copy of the repository on GitHub. This preserves upstream history while granting write access to modify the skill definition. Your fork will eventually replace the upstream entry in your local plugin marketplace.

Step 2: Edit the Canonical Skill File

Open skills/i-have-adhd/SKILL.md and modify the rule set. You can edit existing rules, remove constraints that conflict with your workflow, or append additional guidelines. For example, to add an 11th rule that highlights URLs with an emoji prefix:

### 11. Highlight important URLs

When a response contains a URL, prepend it with "🔗" so the reader can spot links instantly.

Save your changes; the runtime reads this file verbatim at execution time, so any modifications immediately shape the LLM's output style.

Step 3: Synchronize the Cursor Mirror

The repository maintains a duplicate skill file at .cursor/skills/i-have-adhd/SKILL.md for Cursor IDE compatibility. After editing the canonical file, copy your changes to this mirror location to prevent tooling conflicts. While Claude Code and other runtimes read the canonical path, Cursor specifically looks for the file under the .cursor/ directory.

Publish Your Customized Skill

After modifying the rules, redirect your local runtime to load the skill from your fork rather than the upstream source. The exact CLI syntax varies by runtime, but the workflow follows this pattern:

# Remove the upstream skill from your local environment

claude plugin uninstall i-have-adhd
claude plugin marketplace remove i-have-adhd

# Register your fork as the new marketplace source (replace <your-username>)

claude plugin marketplace add <your-username>/i-have-adhd

# Install the skill from your forked repository

claude plugin install i-have-adhd@i-have-adhd

This sequence points the plugin loader to pull skills/i-have-adhd/SKILL.md from your GitHub fork. Subsequent LLM responses will follow your customized rule set instead of the original 10 rules.

Verify Your Custom Rules

Trigger the skill with a test command such as /i-have-adhd or any prompt that generates structured output. If you implemented the URL highlighting rule, inspect the response for the "🔗" prefix before hyperlinks. Successful verification confirms that the runtime is reading your modified SKILL.md from the marketplace entry pointing to your fork.

Summary

• The canonical rule definition lives in skills/i-have-adhd/SKILL.md; editing this file changes how the AI formats responses.
• Maintain parity with .cursor/skills/i-have-adhd/SKILL.md if you use Cursor IDE to avoid configuration drift.
• Republishing requires uninstalling the upstream skill, adding your fork to the marketplace, and reinstalling from the new source.
• The runtime reads markdown rules at execution time, so changes take effect immediately after reinstalling the plugin.

Frequently Asked Questions

Can I add more than 10 rules to the skill?

Yes. The file structure supports any number of rules. Append additional H3 sections to skills/i-have-adhd/SKILL.md following the existing pattern. The runtime applies all detected rules sequentially, regardless of the total count.

Why do I need to update the .cursor/skills/i-have-adhd/SKILL.md mirror?

The mirror exists solely for Cursor IDE compatibility. While Claude Code and other runtimes read the canonical path at skills/i-have-adhd/SKILL.md, Cursor specifically looks for the file under .cursor/. Keeping both synchronized ensures consistent ADHD-friendly formatting across all editing environments.

Will my fork automatically receive upstream updates?

No. Forks are static snapshots. If the original ayghri/i-have-adhd repository updates the default rules, you must manually merge those changes into your fork to benefit from improvements. You may also maintain a divergent rule set indefinitely if the upstream defaults do not suit your needs.

Do I need programming skills to customize the rules?

No coding is required. The skill file uses plain markdown with semantic headings. As long as you can edit text files and use basic GitHub operations, you can customize the 10 ADHD-friendly rules without writing JavaScript, Python, or other programming languages.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}