Where Are Core i-Have-ADHD Skill Definitions and Rules Stored?

The core skill definitions and rules for i-have-adhd are stored in skills/i-have-adhd/SKILL.md, which contains the complete skill specification including metadata, persistence settings, and ten output-shaping rules for ADHD-friendly responses.

The i-have-adhd project is an open-source Agent Skill that helps AI assistants format responses for readers with ADHD. According to the ayghri/i-have-adhd source code, the entire skill specification lives in a single Markdown file following the Agent Skills specification. This architecture makes the skill portable across multiple AI assistant platforms including Cursor and Gemini.

The Central Skill File: skills/i-have-ADHD/SKILL.md

The file skills/i-have-adhd/SKILL.md serves as the authoritative source for how the i-have-adhd skill operates. This file is structured in two parts: a YAML frontmatter block for metadata and a Markdown body for the ruleset.

Frontmatter Metadata

The frontmatter declares critical configuration:

• Skill name: i-have-adhd
• Description: Explains the skill's purpose for ADHD readers
• License: Project licensing information
• disable-model-invocation: true: A persistence flag that prevents the model from invoking itself recursively

The Ten Output-Shaping Rules

The body of SKILL.md enumerates ten specific rules that govern response formatting. These rules include directives such as:

• Lead with the next action — Put the immediate step first
• Number multi-step tasks — Use sequential numbering for clarity
• Give specific time estimates — Provide concrete duration predictions
• Chunk information — Break content into digestible segments

These rules are implemented at lines 1-143 of the source file, making them directly addressable for debugging or extension.

How to Activate the Skill in Practice

The i-have-adhd skill integrates into AI assistants through two primary invocation methods:

Explicit Slash Command

Trigger the skill directly with:

/i-have-adhd

Implicit Domain Matching

The skill also activates automatically when you describe tasks matching its domain:

Add a new auth route that uses JWT and show the updated test results.

When active, the assistant applies all rules from SKILL.md to structure its response.

Example ADHD-Formatted Response

A response following the skill's rules looks like this:

1. Install the JWT package: `npm install jsonwebtoken`.
2. Open `src/auth.ts` and replace the `verifyToken` function (lines 42-58) with the snippet below.
3. Run the auth tests: `npm test -- auth.spec.ts`.

⏱ About 15 minutes if tests already cover the flow.
✅ Login now works with magic links. Try `npm run dev` and open `/login`.

Notice the numbered steps, specific time estimate, and clear action-statement pattern—all mandated by rules in SKILL.md.

Supporting Files in the Repository

While SKILL.md contains the definitive rules, several companion files enable cross-platform distribution:

File	Purpose
skills/i-have-adhd/SKILL.md	Primary skill definition and rule set
README.md	Project overview and quick-start guidance
GEMINI.md	Gemini agent import example: @./skills/i-have-adhd/SKILL.md
.cursor/skills/i-have-adhd/SKILL.md	Cursor IDE integration copy
plugin.json	Open Code plugin declaration for discovery

The Cursor-specific copy in .cursor/skills/i-have-adhd/SKILL.md is a duplicate, not a fork—both files remain synchronized to ensure consistent behavior across editors.

Summary

• The core i-have-adhd skill definitions and rules live exclusively in skills/i-have-adhd/SKILL.md.
• This file combines YAML frontmatter for configuration with Markdown for the ten formatting rules.
• The disable-model-invocation: true flag prevents recursive model calls.
• Platform integrations (Cursor, Gemini) reference this same file, often via duplication or import path.

Frequently Asked Questions

What format does the skill definition file use?

The file follows the Agent Skills specification, using YAML frontmatter for metadata and standard Markdown for the rules body. This hybrid format allows both machine parsing (for activation) and human readability (for editing).

Can I modify the rules for my own use?

Yes. Since SKILL.md is plain Markdown with structured frontmatter, you can edit the rules directly in any text editor. Changes take effect when the skill is reloaded by your AI assistant. The repository is open-source, so fork and customize as needed.

Why is there a duplicate in .cursor/skills/?

Cursor IDE requires skills to reside in a .cursor/ subdirectory for native integration. Rather than restructuring the project, the maintainer placed a duplicate at .cursor/skills/i-have-adhd/SKILL.md. Both copies should remain identical to ensure cross-platform consistency.

How does the skill know when to activate?

Activation depends on the consuming AI assistant's implementation. Most check for explicit slash commands (/i-have-adhd) or pattern-match user prompts against the skill's declared domain. The plugin.json file aids discovery by declaring the repository as an Open Code plugin.

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