Skill i-have-adhd Released to Prevent Answers from Getting Buried

On September 9, 2026, an add-on that changes the response style of coding assistants was released on GitHub. Named i-have-adhd, it is advertised as usable regardless of whether the user has a diagnosis of attention-deficit/hyperactivity disorder. Coverage by ayghri on GitHub Trending highlights its design of presenting the conclusion and next action at the beginning of responses. Its hallmark is eliminating verbose preambles and pleasantries to focus on actionable instructions.

It targets coding assistants that run on the CLI. It can be installed as a skill and plugin for Claude Code. Documentation supports display in English and multiple other languages, with installation steps summarized in AGENTS.md. It is licensed under MIT, permitting forking and modification.

Mechanism That Restructures Responses to Be

Conclusion-First

The purpose of this skill is to prevent answers from getting buried in text. Coverage by ayghri on GitHub Trending outlines the following principle:

Action first. Steps numbered. No “Hope this helps!”

Put action first, number the steps, and omit closing pleasantries. These three points form the core. By prescribing the structure of responses themselves, it aims to reduce the reader’s cognitive load.

The published materials show examples of responses before and after the change. Before the change, the response is a long explanation touching on the authentication flow and dependencies in general. It mentions target files, but the concrete order of operations is unclear. After the change, it is narrowed down to only the installation command, editing targets, verification command, and next branching condition. Open line 42 of src/auth.ts, replace verifyToken, and run auth.spec.ts. If it fails, paste the first failing line. In this form, the reader is not left wondering what to do next.

This contrast is not merely a matter of stylistic preference. In coding work, the clearer the next step is, the easier it is to resume after an interruption. Reducing the burden on working memory directly translates to practical speed. The value lies less in shortening responses themselves than in increasing the identifiability of actions.

Response Style Defined by 10 Rules

The response style is defined by 10 rules contained in SKILL.md. They are as follows. Start by writing the next action. Number multi-step tasks. End with one concrete next step. Curb digressions. Restate the status each time. Show time estimates in minutes. Visualize progress. Treat errors matter-of-factly as facts. Limit lists to a maximum of five items. Do not add formulaic opening, summary, or closing statements.

Restating the status is a device to maintain conversational continuity. It briefly shows the current position and remaining work each time. This spares users from having to scroll back through past messages. Showing estimates in minutes rather than as “a little while” is also practical. It makes it easier to decide whether to start a task and is expected to help prevent procrastination.

The cap on lists is also an important constraint. The more options there are, the slower the decision. Limiting to a maximum of five forces prioritization. The rule to treat errors matter-of-factly prevents the insertion of apologies or excuses. Only descriptions directly linked to cause and remedy remain. This resonates with the discussion of dialogue design covered in Anthropic Declares Fable 5 Revival. It is a stance that prioritizes moving work forward over politeness of responses.

How to Install the Plugin and Customize It

Installation begins by pasting into the CLI input field. The documentation describes installing the skill from the public repository and following the steps in AGENTS.md. In Claude Code, the skill is invoked under the name /i-have-adhd. Calling it after a restart switches the response style.

Procedures for custom adjustments are also published. First, fork the repository and edit skills/i-have-adhd/SKILL.md. Next, remove the existing copy and reinstall from the forked marketplace. As example commands, removal of the copy and removal of the marketplace, addition of the marketplace, and installation of the skill are shown. Care must be taken to avoid name duplication, requiring removal of the upstream copy first. This procedure makes it possible to maintain rules tailored to an organization’s or individual’s preferences.

This fork-based design reflects the idea of treating response style not as a personal setting but as code. It is managed as a document revision rather than an option in a settings screen. Version control and sharing become easier, lending itself to accumulating practical knowledge. This differs from product-side optimization like that shown in Anker Unveils MindBase at IFA 2026, Putting AI at the Core. Its distinctive feature is that the user owns the response behavior.

Positioning in Development Workflows and

Practical Significance

Verbosity in coding assistants has long been an issue. Detailed explanations and background supplements can be useful for beginners. In practice, however, there are many situations where only the next step and how to verify it are needed. Long responses increase scrolling and reading time. They can also cause users to lose track of progress after an interruption.

i-have-adhd addresses this problem through stylistic constraints. It changes the structure of the output rather than the model’s reasoning ability. It maintains the quality of reasoning while reducing transmission loss. Requiring no additional computational resources is also practical.

As a reference, there is a mention of The Adult ADHD Tool Kit by J. Russell Ramsay and Anthony L. Rostain. However, it is not a direct transfer of organizational techniques for humans. It is explicitly stated that they have been adapted to how language models should respond. The focus is on machine output design, not human behavior change. This distinction is important. While bearing the diagnostic name, it does not limit its audience to a specific user group. It is open to any user seeking clarity in work, regardless of cognitive characteristics.

As with ad blocking and efforts to maintain focus, adjustments on the environment side determine productivity. As with the practice shown in Blocking Ads on Android: Combining Brave and DNS Proves Effective, combining tools yields results. Optimizing response style can also be positioned as part of setting up the development environment. Introducing a small skill may reduce friction in daily interactions.

Editorial Opinion

Looking at short-term impact, we expect to see an increase in forks of skills that control coding-assistant responses through style over the next 3–6 months. We assess that short rule documents like SKILL.md will become units of sharing, and response standards for each organization will be version-controlled. We anticipate that a culture prioritizing a verifiable next step over polite long-form text will spread in practical workplaces.

From a long-term perspective, we assess that ownership of output style will shift to the user side. It seems likely to move from a stage of conforming to the product’s default politeness to a stage of choosing a style tailored to job roles and cognitive characteristics. We see restating status and minute-level estimates as encouraging integration with task management. Dialogue logs with the assistant could become progress logs as they are.

As a question from the editorial team, we raise the balance between conciseness and accountability. We see conclusion-first as increasing speed, but omitting rationale risks overlooking errors. We assess that suppressing digressions aids focus but may also reduce opportunities to discover alternatives. What information to cut and what to keep. We believe the mechanism by which users can verify that line is now being tested.

References

• “ayghri /

  i-have-adhd", by **ayghri** — GitHub Trending, 2026-09-09 (ARR)

• Source URL: https://github.com/ayghri/i-have-adhd