How to Configure i-have-adhd for Persistent Output Styling in AGENTS.md

Add the rule block from INSTALL.md to your agent's AGENTS.md file to make the ADHD-friendly output style persistent across all sessions without using the slash command.

The i-have-adhd open-source project by ayghri structures AI outputs to reduce cognitive load for users with ADHD. While the repository provides a temporary /i-have-adhd slash command for single-turn activation, you can configure i-have-adhd for persistent output styling in AGENTS.md to ensure every response follows the action-first format automatically. This approach leverages the agent configuration files that Codex, Zed, Hermes, and Pi read at startup to apply rules globally or per-project.

Understanding the AGENTS.md Mechanism

Agents that support the AGENTS.md specification parse this file during initialization and prepend its instructions to every subsequent prompt. According to the i-have-adhd source code, the repository ships with a ready-to-use markdown block in INSTALL.md that defines the exact output style rules found in skills/i-have-adhd/SKILL.md. When placed in AGENTS.md, these rules persist for the entire session without requiring the /i-have-adhd activation command.

Locating the Configuration Block

The project organizes its ADHD-friendly rules in two key locations. In skills/i-have-adhd/SKILL.md, the skill defines the output constraints that agents enforce when the skill is active. However, for persistent configuration, reference INSTALL.md at lines 89-108, which contains the "Always-on" section with the complete markdown block ready for AGENTS.md insertion.

The block defines ten specific output constraints:

1. Lead with the answer or next action: command, path, or snippet first.
2. Number multi-step work; one bounded action per step.
3. End with one next action doable in under two minutes.
4. Finish the current issue before raising a new one.
5. Restate progress each turn ("step 3 of 5 done").
6. Give time estimates in concrete units, never "a bit".
7. After a change, show what now works.
8. Errors: state location, cause, and fix. No drama.
9. Cap lists at 5 items.
10. No preamble, no recaps, no closers.

Agent-Specific AGENTS.md Locations

Different agents expect the AGENTS.md file in specific directories:

Agent	AGENTS.md Location	Scope
Codex	~/.codex/AGENTS.md	Global (all sessions)
Zed	~/.config/zed/AGENTS.md	Global (all sessions)
Hermes	./AGENTS.md or SOUL.md	Project-local or session
Pi	./AGENTS.md (project root)	Project-local

Configuring Persistent Output Styling in AGENTS.md

For Codex (Global Configuration)

Create the configuration directory if it doesn't exist, then append the output style block:

mkdir -p ~/.codex
cat >> ~/.codex/AGENTS.md <<'EOF'

## Output style

The reader has ADHD. Shape every response so it can be acted on:

1. Lead with the answer or next action: command, path, or snippet first.
2. Number multi-step work; one bounded action per step.
3. End with one next action doable in under two minutes.
4. Finish the current issue before raising a new one.
5. Restate progress each turn ("step 3 of 5 done").
6. Give time estimates in concrete units, never "a bit".
7. After a change, show what now works.
8. Errors: state location, cause, and fix. No drama.
9. Cap lists at 5 items.
10. No preamble, no recaps, no closers.

*Exceptions:* explain fully when asked to explain. Confirm before destructive actions. After three failed fixes, stop and name the doubtful assumption. If the request is ambiguous, ask one short question.
EOF

For Project-Specific Agents (Hermes/Pi)

Navigate to your project root and create the AGENTS.md file:

cat >> AGENTS.md <<'EOF'

## Output style

The reader has ADHD. Shape every response so it can be acted on:

1. Lead with the answer or next action: command, path, or snippet first.
2. Number multi-step work; one bounded action per step.
3. End with one next action doable in under two minutes.
4. Finish the current issue before raising a new one.
5. Restate progress each turn ("step 3 of 5 done").
6. Give time estimates in concrete units, never "a bit".
7. After a change, show what now works.
8. Errors: state location, cause, and fix. No drama.
9. Cap lists at 5 items.
10. No preamble, no recaps, no closers.

*Exceptions:* explain fully when asked to explain. Confirm before destructive actions. After three failed fixes, stop and name the doubtful assumption. If the request is ambiguous, ask one short question.
EOF

Verifying the Configuration

After saving the file, restart your agent completely to trigger the configuration reload. Test by requesting a multi-step explanation; if configured correctly, the response will begin with the immediate action or answer, followed by numbered steps under five items, and conclude with a single next action estimate. If the agent still requires the /i-have-adhd command, verify the file path matches your specific agent's expected location from the table above.

Summary

• Copy the output style block from INSTALL.md lines 89-108 to make the skill persistent
• Place AGENTS.md in ~/.codex/ for Codex or ~/.config/zed/ for Zed global configuration
• Use project-root AGENTS.md for Hermes and Pi per-project styling
• Restart the agent after creating the file to load the rules
• The configuration eliminates the need for manual /i-have-adhd activation

Frequently Asked Questions

Where do I find the exact text to copy into AGENTS.md?

The exact markdown block resides in INSTALL.md at lines 89-108 in the ayghri/i-have-adhd repository. This section contains the same ten output style rules defined in skills/i-have-adhd/SKILL.md, formatted specifically for AGENTS.md insertion.

Does AGENTS.md work with all AI agents?

No, AGENTS.md support varies by agent. According to the repository documentation, Codex, Zed, Hermes, and Pi explicitly support AGENTS.md configuration files. Other agents may require different configuration methods or the temporary /i-have-adhd slash command.

Can I use AGENTS.md for just one project instead of globally?

Yes. Agents like Hermes and Pi support project-specific AGENTS.md files in the working directory or project root. For Codex and Zed, you typically configure global settings, but you can override these per-project if the agent supports workspace-specific configuration files.

What happens if I still type /i-have-adhd after configuring AGENTS.md?

The /i-have-adhd slash command remains functional and will apply the same rules, but it becomes redundant since the AGENTS.md configuration already prepends these instructions to every prompt. Using both simultaneously does not cause conflicts but offers no additional benefit.

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