# Rule 10 Forbidden Elements in i-have-adhd: No Preamble, No Recap, No Closers

> Discover Rule 10 forbidden elements in i-have-adhd. Learn how to eliminate preambles recaps and closers for direct actionable AI assistance. Maximize efficiency now.

- Repository: [Ayoub Ghriss/i-have-adhd](https://github.com/ayghri/i-have-adhd)
- Tags: best-practices
- Published: 2026-08-21

---

**Rule 10 prohibits all conversational openers, mid-response recaps, and closing pleasantries to ensure every assistant response delivers immediate, actionable content without distraction.**

The `ayghri/i-have-adhd` repository enforces a strict communication framework designed specifically for readers with ADHD. Rule 10 (also designated as Rule IO) defines the forbidden elements that prevent assistants from using soft language or redundant framing that could interrupt focus or delay critical information.

## The Three Categories of Forbidden Elements

Rule 10 splits prohibited language into three distinct categories. These restrictions are defined verbatim in [`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md) at lines 109‑113.

### Openers (Preamble)

**Openers** include any introductory phrase that previews, explains, or apologizes before delivering the actual answer. These preamble statements create unnecessary cognitive load by delaying the core content.

Forbidden examples include:

- "Great question"
- "Let me..."
- "Sure!"
- "Looking at your..."
- "To answer your question..."

### Recaps (Step Summaries)

**Recaps** are summaries that repeat actions already completed, particularly those ending with statements like "I've now done X, Y, and Z..." These mid-response recaps disrupt workflow by stating information the user has already processed.

Forbidden examples include:

- "I've now done X, Y, and Z, which means..."
- "We're now finished, next..."

### Closers (Closing Pleasantries)

**Closers** encompass any concluding sentence that invites further interaction or offers well-wishes. These phrases add no informational value and signal a premature end to the task at hand.

Forbidden examples include:

- "Let me know if you need anything else"
- "Hope this helps"
- "Happy to clarify"
- "Feel free to ask"

## Source Code Location and Implementation

The complete rule set resides in the skill definition file. According to the `ayghri/i-have-adhd` source code, the specific prohibitions appear in the following location:

- **File:** [`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md)
- **Lines:** 109‑113

By enforcing these constraints at the skill level, the repository ensures that assistant responses jump straight to actionable content, maintain concise structure, and avoid any language patterns that could cause readers to lose focus.

## Compliant vs. Non-Compliant Response Examples

The following examples illustrate the practical application of Rule 10.

**Compliant response:**

```markdown
Run `npm install jsonwebtoken`, then edit `src/auth.ts:42`.

```

**Non-compliant preamble:**

```markdown
Let me show you how to fix this issue…

```

**Non-compliant recap:**

```markdown
I've now updated the auth file, ran the tests, and everything works.

```

**Non-compliant closer:**

```markdown
Hope this helps! Let me know if you need anything else.

```

In production implementations, the assistant filters out any patterns matching the three forbidden categories before returning the final response.

## Summary

- **Rule 10 (Rule IO)** eliminates three types of conversational overhead: openers, recaps, and closers.
- **Openers** are introductory phrases like "Great question" or "Let me..." that delay content delivery.
- **Recaps** are redundant summaries of completed steps that restate obvious progress.
- **Closers** are pleasantries like "Hope this helps" that add no actionable value.
- The rule **source** is located at lines 109‑113 in [`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md).
- **Implementation** requires filtering all assistant outputs to ensure immediate, distraction-free information delivery.

## Frequently Asked Questions

### What is the difference between Rule 10 and Rule IO?

Rule 10 and Rule IO refer to the same rule in the `i-have-adhd` skill. The designation "Rule IO" appears in the source documentation, while "Rule 10" is the common ordinal reference. Both prohibit the same three elements: preambles, recaps, and closers.

### Why are recaps forbidden if they help summarize information?

Recaps force readers with ADHD to re-process information they have already seen, creating cognitive friction. According to the skill definition, summaries that state "I've now done X, Y, and Z..." interrupt workflow by stating the obvious rather than advancing toward the solution.

### Where exactly is Rule 10 defined in the repository?

Rule 10 is defined in [`skills/i-have-adhd/SKILL.md`](https://github.com/ayghri/i-have-adhd/blob/main/skills/i-have-adhd/SKILL.md) at lines 109‑113. This file contains the complete skill definition used by runtimes including Claude, Codex, and OpenCode to filter assistant responses.

### Can assistants use any closing phrase at all under Rule 10?

No. Rule 10 explicitly prohibits all closing pleasantries, including "Let me know if you need anything else," "Happy to clarify," and "Feel free to ask." The only permissible content is actionable information directly responding to the user's need.