Why Is i-have-adhd Not Appearing in Slash Command Autocomplete? Troubleshooting Guide

If the /i-have-adhd slash command is missing from autocomplete, verify the skill is installed in an agent-scanned directory, ensure enableSkillCommands is set to true in your agent's settings.json, and restart your session.

The ayghri/i-have-adhd repository provides a specialized skill that enhances AI coding assistants with ADHD-aware prompting patterns. When properly configured, typing /i-have-adhd triggers an autocomplete suggestion that injects contextually relevant rules from skills/i-have-adhd/SKILL.md into your session.

How the i-have-adhd Slash Command Gets Registered

Five distinct components must align for the autocomplete entry to surface. According to the source code in ayghri/i-have-adhd, the registration pipeline involves:

Skill Definition File

The core identity resides in skills/i-have-adhd/SKILL.md. This file declares the command name (i-have-adhd) and defines the rule set injected upon invocation. Without this file in the correct location, the agent cannot recognize the command.

Skill Registration Location

Agents scan specific directories for available skills. For Pi, these include ~/.pi/agent/skills and ~/.agents/skills (global), plus project-local paths like .pi/skills and .agents/skills. The i-have-adhd folder must exist in one of these scanned paths.

Slash-Command Enable Flag

Even with files in place, agents filter out skill commands unless explicitly enabled. The settings.json configuration must contain "enableSkillCommands": true. This flag is referenced at line 254 in INSTALL.md as the primary gatekeeper for slash-command visibility.

Session Start Hook (Always-On Mode)

The optional hooks/always-on.sh script automatically injects the skill rules when it detects the flag file $CLAUDE_CONFIG_DIR/.i-have-adhd-always. This bypasses manual command invocation but requires the flag file to exist.

Plugin Manifest

The plugin.json file at the repository root declares that this package provides a skill named i-have-adhd, allowing agent-specific tooling to index the command.

Common Causes for i-have-adhd Missing from Autocomplete

Skill Installed Outside Discovered Paths

If you copied the skill folder to a non-standard directory (e.g., /usr/local/skills instead of ~/.pi/agent/skills), the agent's filesystem scanner never indexes it. Agents strictly search their predefined skill directories.

Disabled Slash-Command Flag

Default agent configurations often ship with "enableSkillCommands": false. Without toggling this setting, the UI actively filters out all / prefixed skill commands from the autocomplete menu.

Stale Session State

Configuration changes require a fresh session. The agent parses settings.json and scans skill directories only at startup. Editing files while a session is active will not update the autocomplete list until you restart.

Missing Always-On Flag File

When using the always-on hook for Claude Code, the presence of ~/.claude/.i-have-adhd-always is mandatory. If this file is absent, hooks/always-on.sh exits silently without loading the skill context.

Step-by-Step Resolution Checklist

Follow these verification steps to restore the /i-have-adhd autocomplete entry:

1. Verify installation path. Confirm the i-have-adhd folder exists in your agent's skill directory (e.g., ~/.pi/agent/skills/i-have-adhd).

2. Enable skill commands. Open your agent's settings.json and set the boolean flag:

   {
     "enableSkillCommands": true
   }

3. Restart the agent. Close and reopen your chat session, or reload the agent configuration to trigger a fresh skill scan.

4. Confirm registration. Use the command-line tool to verify the skill is indexed:

   npx skills list | grep i-have-adhd

   Expected output: i-have-adhd.

5. Test autocomplete. Type / in the chat interface and verify i-have-adhd appears in the suggestion dropdown.

Agent-Specific Configuration Examples

Pi (Personal Intelligence)

Install via the skills CLI and explicitly enable the feature flag:

# Install to Pi's skill directory

npx skills add ayghri/i-have-adhd -a pi -y

# Enable slash commands in Pi's configuration

cat > ~/.pi/settings.json <<'EOF'
{
  "enableSkillCommands": true
}
EOF

# Restart Pi session, then verify

npx skills list | grep i-have-adhd

Claude Code (Always-On Mode)

For automatic injection without typing the slash command:

# Create the flag file that triggers hooks/always-on.sh

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

# Restart Claude Code

# The skill rules will be active immediately in new sessions

Gemini CLI

Manual command descriptor installation:

# Create Gemini commands directory if missing

mkdir -p ~/.gemini/commands

# Download the command definition from skills/i-have-adhd/agents/gemini.toml

curl -fsSL https://raw.githubusercontent.com/ayghri/i-have-adhd/main/skills/i-have-adhd/agents/gemini.toml \
  -o ~/.gemini/commands/i-have-adhd.toml

# Start a new Gemini session

Summary

• The slash-command autocomplete requires the skill to be installed in an agent-scanned directory like ~/.pi/agent/skills.
• The enableSkillCommands flag in settings.json must be set to true or the agent hides all skill commands.
• Session restarts are mandatory after installing skills or changing configuration files.
• The always-on hook (hooks/always-on.sh) requires the presence of ~/.claude/.i-have-adhd-always to function.
• Verify installation using npx skills list before testing autocomplete.

Frequently Asked Questions

Why doesn't i-have-adhd appear after I installed it?

The skill likely exists outside the agent's search path or the enableSkillCommands flag remains disabled. Verify the folder is in ~/.pi/agent/skills (or equivalent for your agent) and check that settings.json contains "enableSkillCommands": true.

What is the difference between installing the skill and using always-on mode?

Installing the skill via npx skills add makes the /i-have-adhd command available in autocomplete, requiring manual invocation. Always-on mode (triggered by ~/.claude/.i-have-adhd-always) automatically injects the skill rules at session start without typing the command, but still requires the skill files to be present.

Where is the settings.json file located?

The location varies by agent: Pi uses ~/.pi/settings.json, while Claude Code and others use their respective config directories (e.g., ~/.claude/). Refer to your agent's documentation, but the structure always requires the enableSkillCommands boolean at the root level.

Do I need to restart my terminal or just the chat session?

You must restart the agent process or start a new chat session. Simply restarting the terminal window is insufficient if the agent process persists in the background. The agent reads settings.json and scans skill directories only during initialization.

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