i-have-adhd Preamble and Closer Prohibitions: The No-Filler Communication Rules

The i-have-adhd skill enforces three strict prohibitions—no preamble, no recap, and no closing pleasantries—to ensure every response begins immediately with actionable content and terminates when the technical answer is delivered.

The i-have-adhd repository by ayghri defines a specialized communication framework optimized for readers with ADHD. These preamble and closer prohibitions eliminate cognitive friction by removing filler text, social pleasantries, and redundant summaries, allowing users to act on information without navigating unnecessary linguistic overhead.

The Three Core Prohibitions

The framework defines three absolute constraints that shape every interaction. According to Rule 10 in skills/i-have-adhd/SKILL.md (lines 109-116), these rules appear verbatim as: "No preamble, no recap, no closing pleasantries."

No Preamble

Responses must start directly with the actionable output—whether a command, code snippet, or file path. Introductory phrases like "Great question," "Let me explain," or "Sure!" are strictly forbidden. The first character should be the beginning of the solution, not social acknowledgment.

No Recap

After delivering information, the model must not summarize what was accomplished. The response ends when the technical work finishes, without phrases like "I've now updated the schema..." or "We've completed steps 1, 2, and 3." This prevents redundant processing for readers who already tracked the progress.

No Closing Pleasantries

Courteous endings such as "Hope this helps," "Feel free to ask," or "Let me know if you need anything else" are prohibited. The final line must contain the last concrete action or piece of information, stopping immediately when the value delivery ends.

Source Code Implementation

These prohibitions are hardcoded in the repository's skill definition files. The same wording is reproduced in skills/i-have-adhd/agents/gemini.toml (lines 19-20), where the Gemini CLI command definition enforces these rules for the /i-have-adhd shortcut. This dual definition ensures consistency whether the skill is invoked through the markdown protocol or the Gemini agent configuration.

Practical Examples: Allowed vs. Forbidden Output

The following examples demonstrate how the preamble and closer prohibitions apply in practice according to the skill specifications.

Command Delivery

Allowed:

npm install jsonwebtoken

Forbidden:

Let's install the package. Run `npm install jsonwebtoken`.

Task Completion

Allowed:

Next: run `npm test` and paste the first failing line.

Forbidden:

That's it! Hope this helps. Let me know if you have more questions.

Step-by-Step Instructions

Allowed:

1. Open src/auth.ts
2. Replace verifyToken...
3. Run npm test

Forbidden:

Here's what we'll do: first we'll open the file, then we'll replace the function, and finally we'll test it.

Error Reporting

Allowed:

Test fails at auth.spec.ts:42: expected 200, got 401. Fix: add Authorization header.

Forbidden:

Uh oh, the test is failing. There seems to be a problem...

Summary

• The i-have-adhd skill enforces three strict communication prohibitions: no preamble, no recap, and no closing pleasantries.
• These rules are defined in skills/i-have-adhd/SKILL.md (Rule 10, lines 109-116) and skills/i-have-adhd/agents/gemini.toml (lines 19-20).
• Responses must begin immediately with actionable content (commands, paths, or code) and end with the final concrete action.
• This structure minimizes cognitive load for readers with ADHD by eliminating filler text and social boilerplate that interrupts task execution.

Frequently Asked Questions

What specific phrases violate the no-preamble rule?

Any introductory text that delays the actionable content violates this rule. Examples include "Great question," "Let me explain," "Sure, I can help with that," or "Looking at your request..." The response must begin with the command, code, or path the user needs, starting at line 1, column 1 with the technical solution.

Why does the skill prohibit recaps if they might help with memory retention?

While recaps can aid retention for some readers, they add verbal overhead that increases cognitive load for ADHD readers who have already processed the steps. The skill assumes the user can scroll up to review previous actions if needed. By ending immediately when the work is done, the response respects the user's working memory capacity and prevents redundant information processing.

How are these rules technically enforced in the codebase?

The prohibitions are embedded as system prompt constraints in skills/i-have-adhd/SKILL.md and the Gemini command configuration at skills/i-have-adhd/agents/gemini.toml. When the /i-have-adhd shortcut is invoked, the model receives these constraints as part of its instruction set, compelling generation that complies with the no-preamble, no-recap, and no-pleasantries requirements.

Can these rules be adapted for other communication contexts?

Yes. The framework defined in the i-have-adhd repository can be applied to any technical documentation or AI interaction design. The specific constraint phrasing used in skills/i-have-adhd/SKILL.md serves as a template for creating ADHD-friendly communication protocols in developer tools, documentation systems, or customer support workflows where immediate actionability is prioritized.

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