How to Access the i-have-adhd API Endpoints in Claude Code and Codex

The i-have-adhd repository exposes skill-based "API endpoints" through slash commands in Claude Code (/i-have-adhd) and dollar-prefixed shortcuts in Codex ($i-have-adhd), requiring no authentication tokens or HTTP requests.

The i-have-adhd repository by ayghri does not provide a traditional REST API. Instead, it functions as a skill plugin that reformats LLM outputs into ADHD-friendly structures through command-line invocations within Claude Code and Codex agents. Accessing these endpoints means installing the skill and triggering its output-shaping behavior through specific chat commands.

Understanding the i-have-adhd API Architecture

Unlike conventional web APIs that require HTTP requests and authentication headers, this repository implements a skill-based interface that operates entirely within LLM-powered coding assistants. The "endpoints" are command invocations that process the LLM's generated response and return reformatted, ADHD-friendly versions featuring action-first formatting, numbered steps, and eliminated tangents.

The skill enforces 10 formatting rules defined in skills/i-have-adhd/SKILL.md, which transform verbose LLM outputs into concise, structured responses optimized for ADHD readers.

Installing the Skill in Claude Code

To access the i-have-adhd endpoint in Claude Code, install the plugin from the marketplace. No local clone of the repository is required, as Claude Code fetches the skill automatically from the remote source.

Use the following commands to install and activate the skill:

# Add the skill from the marketplace

claude plugin marketplace add ayghri/i-have-adhd

# Install the plugin at the latest commit (main)

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

Once installed, invoke the skill within any Claude Code chat session by typing:

/i-have-adhd

According to the repository's README.md, this slash command triggers the skill's output-shaping behavior immediately.

Installing the Skill in Codex

For Codex users, installation follows a similar pattern using the Codex CLI. The skill can be pinned to the main branch to ensure stability.

Install the plugin using these commands:

# Add the skill from the marketplace (pin to the main branch)

codex plugin marketplace add ayghri/i-have-adhd --ref main

# Add the plugin to the current workspace

codex plugin add i-have-adhd@i-have-adhd

Invoking the i-have-adhd Endpoints

After installation, accessing the API endpoints requires specific invocation commands depending on your agent.

Explicit Invocation

Claude Code: Use the slash command /i-have-adhd in the chat window to explicitly apply the ADHD-friendly formatting to the LLM's next response.

Codex: Prepend your prompt or code block with the dollar-prefixed shortcut $i-have-adhd to trigger the reformatting explicitly.

Implicit Invocation

Both agents support implicit invocation when the LLM detects a multi-step answer that could benefit from the ADHD-friendly format. The skill automatically applies its 10 formatting rules from SKILL.md without requiring explicit commands when the context suggests structured output would improve clarity.

Key Configuration Files

The i-have-adhd skill relies on specific metadata files to define its behavior across platforms:

• skills/i-have-adhd/SKILL.md – Defines the 10 formatting rules the skill enforces, including action-first structure and numbered step conventions.
• .claude-plugin/plugin.json – Metadata file used by Claude Code to load and validate the skill.
• .codex-plugin/plugin.json – Metadata file used by Codex to load and validate the skill.
• README.md – Contains installation instructions and usage commands for both Claude Code and Codex.

These files work together to implement the skill's behavior; no external API endpoints or network services are required beyond the initial plugin installation.

Summary

• The i-have-adhd repository provides skill-based API endpoints rather than traditional HTTP APIs.
• Claude Code users install via claude plugin marketplace add and invoke with /i-have-adhd.
• Codex users install via codex plugin marketplace add --ref main and invoke with $i-have-adhd.
• The skill automatically reformats LLM outputs according to rules defined in SKILL.md, creating action-first, numbered responses without tangents.
• No authentication tokens, API keys, or local server setup is required to access these endpoints.

Frequently Asked Questions

Does the i-have-adhd repository expose REST API endpoints?

No. The repository does not expose traditional REST or HTTP API endpoints. Instead, it functions as a skill plugin for Claude Code and Codex, where "endpoint access" means invoking slash commands (/i-have-adhd) or dollar-prefixed shortcuts ($i-have-adhd) within the chat interface.

Do I need authentication credentials to access the i-have-adhd API?

No authentication tokens, API keys, or credentials are required. The skill operates entirely within the LLM agent's environment after installation. Claude Code and Codex handle all necessary permissions through their respective plugin marketplaces during the installation process.

Can I use the i-have-adhd skill without installing it in Claude Code or Codex?

No. The skill requires installation through the official plugin marketplaces to function. However, once installed, it can operate in implicit mode, automatically applying ADHD-friendly formatting when the LLM detects suitable content, without requiring explicit command invocation for every interaction.

Where are the formatting rules for the i-have-adhd API defined?

The 10 formatting rules that govern the skill's output-shaping behavior are defined in skills/i-have-adhd/SKILL.md. This file specifies the action-first structure, numbered step requirements, and elimination of tangential content that characterizes the ADHD-friendly output format.

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