What Are the Two Activation Modes for the i-have-adhd Skill?

The i-have-adhd skill supports two activation modes: explicit invocation via slash commands for temporary per-session use, and always-on persistent mode via configuration files for automatic application at startup.

The ayghri/i-have-adhd repository provides an AI assistant skill designed to format responses for ADHD-friendly readability. According to the installation documentation in INSTALL.md, users can activate these output formatting rules either on-demand or permanently, depending on their workflow requirements.

Explicit Invocation Mode

Explicit invocation activates the skill only when you manually trigger it during a specific session. This mode is ideal for ad-hoc tasks where you need ADHD-optimized formatting temporarily.

In Claude Code or Qwen Code, type the following command to activate the skill for the current session:

/i-have-adhd

For Codex-compatible environments, use the dollar-sign prefix instead:

$i-have-adhd

Once activated, the rules defined in skills/i-have-adhd/SKILL.md remain active until you explicitly disable them by saying "stop adhd mode" or "normal mode". This ensures the formatting persists across multiple turns in your conversation without affecting other sessions.

Always-On Persistent Mode

Always-on mode automatically loads the i-have-adhd rules at the start of every session, eliminating the need to type the activation command repeatedly.

For Claude Code users, create an empty marker file in your home directory:

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

Claude Code checks for this file at session startup and automatically loads the skill rules when present.

For other AI harnesses like Cursor or OpenCode, paste the always-on snippet from INSTALL.md into your agent's persistent rules configuration. For example, in Cursor, add the rules to Settings → Rules → User Rules or place them in .cursor/rules/. For OpenCode, add them to ~/.config/opencode/AGENTS.md. After saving the configuration, restart your agent to apply the changes.

Choosing the Right Activation Method

Use explicit invocation when you need ADHD-optimized formatting occasionally or want to preserve default behavior for most interactions. Use always-on mode when you require the structured, step-by-step output format for every session, ensuring consistency without manual intervention.

Summary

• The i-have-adhd skill offers two activation modes: explicit invocation and always-on persistent mode.
• Explicit mode uses /i-have-adhd (Claude/Qwen) or $i-have-adhd (Codex) and lasts until you say "stop adhd mode".
• Always-on mode for Claude Code requires creating ~/.claude/.i-have-adhd-always to trigger automatic loading at session start.
• Other harnesses require pasting the skill snippet into persistent rule files like Cursor's user rules or OpenCode's AGENTS.md.
• Both modes reference the same core rules defined in skills/i-have-adhd/SKILL.md.

Frequently Asked Questions

How do I temporarily disable the i-have-adhd skill during a session?

To disable the skill during an active session where you used explicit invocation, simply tell the agent "stop adhd mode" or "normal mode". This immediately reverts the output formatting to standard behavior without affecting future sessions.

Does the always-on mode affect all AI agents or only specific ones?

Always-on mode is specific to the harness you configure. Creating ~/.claude/.i-have-adhd-always only affects Claude Code sessions. For Cursor, OpenCode, or other agents, you must paste the always-on snippet into that specific tool's persistent rules file, as documented in INSTALL.md.

Where are the core rules for the i-have-adhd skill defined?

The core formatting rules—specifying how to structure numbered lists, bold key terms, and break down complex tasks—are defined in skills/i-have-adhd/SKILL.md. Both activation modes ultimately load and apply these 10 specific ADHD-friendly output rules.

Can I use both activation modes simultaneously?

Yes, but doing so is redundant. If you have already enabled always-on mode via the marker file or persistent rules, the skill loads automatically; typing the explicit command is unnecessary. However, using both causes no conflicts since they reference the same underlying rule set.

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