Claude ADHD Plugin: What Exists and How to Install One Anthropic has no official Claude ADHD plugin, but four community GitHub projects share the name and serve different purposes, according to a blog post from the Neurodivergent Systems series. The projects include i-have-adhd (ten output-formatting rules), claude-adhd-skills by ravila4 (time awareness and Obsidian journal), three skills from the Citability repo (mirror, librarian, standing), and a divergent-reasoning tool named adhd. The post explains that Claude Code extensions come as skills or plugins, with plugins installed via marketplace commands, and advises choosing based on the specific gap each project addresses. Claude ADHD Plugin: What Exists and How to Install One There is no official Claude ADHD plugin. Four community projects share the name and do different things. What each changes, and how to install one. Why this matters There is no official Claude ADHD plugin from Anthropic. Four community projects share the name: i-have-adhd ten output-formatting rules, installs as a plugin from a marketplace , claude-adhd-skills by ravila4 time awareness, nudges, and an Obsidian journal, installed by copying folders , the three skills I run from the Citability repo mirror, librarian, standing, which fetch evidence before answering , and a project literally named adhd that is a divergent-reasoning tool with nothing to do with the condition. Pick by the gap you have, not by the name. In this cluster Cluster context This article sits inside Neurodivergent Systems. Open topic hub /topics/neurodivergent-systems ADHD traits as engineering advantages: pattern recognition, resilience, and system design. Most productivity advice assumes a brain that prefers linearity and uniform energy. This cluster is the counter-program: use neurodivergent traits as a design constraint and build systems around them. GTD for ADHD Doesn't Work: The Energy-Based System I Built Instead GTD doesn't work for ADHD brains. The energy-aware productivity system I built instead, hyperfocus scheduling, AI processing, and the workflow I use to ship. I Have 73 Browser Tabs Open. ADHD Made Me a Better Architect. Five ADHD cognitive traits mapped to real systems architecture patterns: pattern recognition, parallel processing, novelty seeking, and chaos resilience. Claude for ADHD: There's No Plugin, Here's My Workflow No, Claude has no built-in ADHD mode or plugin. The 5-step workflow I use instead to beat task-initiation paralysis and context-switching. Part of my Neurodivergent Systems /topics/neurodivergent-systems series on tools and workflows for technical work. Short answer: there is no official Claude ADHD plugin. Anthropic ships no such feature. What the search finds is four community projects on GitHub that share the word ADHD and do four different things. This page sorts them by the gap each one closes, shows how the plugin system installs them, and says which one I run. If you want the install command for my three skills and nothing else, it is on the install page /blog/claude-code-adhd-skills-install . This page is for the reader who typed “claude adhd plugin” and wants to know what they are choosing between. All of it sits inside the ADHD engineer’s productivity system /blog/adhd-engineer-productivity-system , which needs no plugin at all. First, what a plugin is Claude Code has two units of extension, and the word “plugin” gets used for both. A skill is a folder with a SKILL.md file. The file tells Claude when to use the skill and what to do. You invoke it with a slash command or Claude picks it up on its own. Skills live in ~/.claude/skills/ and you can install one by copying the folder there. A plugin is a package. Per the Claude Code documentation, a plugin directory can hold skills, agents, event scripts that run when Claude does things, and tool-server definitions, with an optional manifest at .claude-plugin/plugin.json . Plugins install from a marketplace, which is a GitHub repository or a local folder that lists them: /plugin marketplace add owner/repo /plugin install plugin-name@marketplace-name The same two steps work from the shell as claude plugin marketplace add and claude plugin install . The practical difference: a skill you copy, a plugin you install and uninstall by name. Every ADHD project below is mostly skills. Two of them also package themselves as plugins so the install is one command. That is the only sense in which any of them is “a plugin”. The four projects I read all four READMEs and the skill files on 2026-09-05. The one I run daily is the third. For the others I describe what the code does, not how it felt to use for a month, because I have not done that. 1. i-have-adhd, by ayghri What it changes: how Claude writes to you. Ten rules in one SKILL.md: lead with the next action, number multi-step tasks, end with one concrete next step, suppress tangents, restate state every turn, give time estimates in minutes, make wins visible, report errors flatly, cap lists at five items, no preamble and no closers. The gap it closes: the buried answer. If Claude opens with “Great question ” and puts the command in paragraph four, this is the fix. The README’s before-and-after is a paragraph of throat-clearing turned into a three-line numbered list with the command first. Install: it ships as a plugin. The README gives a one-line prompt you paste into Claude Code that tells it to install from the repository, or the marketplace route: claude plugin marketplace add ayghri/i-have-adhd claude plugin install i-have-adhd@i-have-adhd Then restart Claude Code and invoke /i-have-adhd . Does not do: anything about what Claude does before it answers. It does not read your git history, does not track time, does not remember yesterday. It is a communication layer, and its README says it is loosely based on a clinical toolkit adapted for how a model should respond. Note on the name: this is the project people usually mean by “the i have adhd skill”. It has been translated into six languages, which tells you how far it has spread. 2. claude-adhd-skills, by ravila4 What it changes: time and memory. Four skills and a set of event scripts. A date script injects the current time into every prompt so Claude knows when it is. A nudge system fires timed reminders on prompt submission: “stop me at 11”, “remind me about standup in thirty minutes”. A daily-journal skill asks about your day and writes it to an Obsidian vault. An obsidian-vault skill manages notes so a new session can read yesterday’s entry and catch up instead of you re-explaining. The gap it closes: time blindness and the morning rebuild. Claude has no clock. This project gives it one, then uses the clock to make “wrap up in thirty minutes” stick. Install: clone, copy the skills folder into ~/.claude/skills/ , copy the event scripts into ~/.claude/hooks/ , and paste the event configuration into your settings file. Three steps, not one, and the reminders only work with the settings change. Does not do: change output formatting. Its author’s CLAUDE.md template covers break suggestions and pacing, which is a different layer again. Depends on: Obsidian for the memory half. Without a vault the journal has nowhere to go. The nudges work on their own. 3. The three skills I run: mirror, librarian, standing What it changes: what Claude does before it answers. These three are copied unedited from my own skills directory into github.com/Citability/claude-code-adhd-skills https://github.com/Citability/claude-code-adhd-skills under MIT. Each one goes and gets evidence first. mirror reads the git history when my brain says I did nothing today, and reports what actually shipped. standing reads the week out of session transcripts and git, counts the invisible work, and places it against my own trailing month instead of an imagined peer who ships every day. librarian walks a project knowledge graph and hands back one concrete change when the whole codebase will not fit in my head. The gap it closes: discounting the positive, comparing to a phantom, and working-memory overload. Three executive-function gaps that the other projects do not touch. Install: one line pasted into Claude Code, or clone and copy the three folders. The full walkthrough with the troubleshooting is on the install page /blog/claude-code-adhd-skills-install . Does not do: anything about tone, length, or formatting. librarian needs a docs/codex/ directory in the repository and will tell you and stop if there is none. standing compares you to your field only if you give it a cited benchmark file; otherwise it compares you to your own last month and says so. None of the three has been evaluated beyond the transcripts in the README. Built for this brain ADHD Engineer's Productivity System $19 Notion templates built for how ADHD brains actually work: energy-based task management, two-axis filtering, and a brain-dump capture system. Get the system https://chudinnorukam.lemonsqueezy.com/checkout/buy/f734b2a4-f3cd-41ba-937b-9cb140f847e0 4. adhd, by UditAkhourii What it changes: how Claude reasons on open-ended questions. Despite the name, this is not an ADHD support tool. The README describes it as an architectural fix for premature convergence: it spawns several isolated reasoning processes under deliberately distorted frames, then a critic pass scores and prunes them. It is aimed at design decisions, naming, and “give me a few ways to” prompts. The gap it closes: a model anchoring on its first idea. That is a real problem. It is not an executive-function problem. Why it is here: it ranks for the query, and if you install it expecting help with focus you will be confused. The name is a joke about divergent thinking. Read it as that. Which one for which gap | Your problem | Install | Type | Install effort | |---|---|---|---| | The answer is buried under preamble | i-have-adhd | plugin, one command | one line | | Claude does not know what time it is, and neither do you | claude-adhd-skills | skills plus event scripts | clone, copy, edit settings | | You shipped things and your brain says you did not | mirror and standing | skills | one line | | The codebase will not fit in your head | librarian | skill, needs docs/codex/ | one line, plus the directory | | Claude commits to its first idea too fast | adhd UditAkhourii | npm package | not an ADHD tool | They do not conflict. i-have-adhd shapes the writing. The other two shape the fetching. I run the three Citability skills and nothing from the first two, which is a fact about my gaps, not a ranking. How to tune or fork one Every project above is a folder of text files. That is the whole point of the skill format, and it is the reason “install one and live with it” is the wrong frame. You will change it. Read the SKILL.md first. It is short, usually under a page, and it is the entire behavior. i-have-adhd is ten numbered rules. If rule seven “cap lists at five items” fights the way you work, delete the line. There is no rebuild step. The next session reads the edited file. Change the numbers to yours. My checkpoint interval is 45 minutes because that is the size of task I finish. ravila4’s nudge defaults and my standing comparison window are both parameters, not laws. A skill tuned to its author’s numbers is a skill that quietly stops fitting you, and you will blame yourself rather than the number. Fork rather than edit in place if you installed as a plugin. A plugin installed from a marketplace gets overwritten on update. Per the Claude Code documentation, the plugin’s contents live under its own directory with an optional manifest at .claude-plugin/plugin.json and the skills under skills/