Your brain works differently. Your tools should too.
Hyperfocus restructures Claude's output so your brain can lock in. Not fewer words — better structure.
claude plugin marketplace add nextor2k/hyperfocus
claude plugin install hyperfocus@hyperfocusThen type /hyperfocus in any conversation. Done.
Hyperfocus changes how Claude talks to you.
Every response gets:
- Answer first — key point up top, context after
- Clear sections — subheadings you can scan in 2 seconds
- Short chunks — one idea per paragraph, short sentences
- Lists over walls — bullet points instead of dense prose
Important
Same depth. Same technical accuracy. Just structured so your brain can absorb it.
Pick your level of structure. Switch anytime with /hyperfocus clean|flow|zen.
For: quick answers, PR summaries, terse professional prose.
Shorter paragraphs, front-loaded points, bullet lists. Professional tone stays intact.
For: learning, debugging, technical explanations.
Each section follows What → Why → How. Subheadings every 2–3 paragraphs for easy re-entry after breaks.
For: dense docs, architecture discussions, long sessions.
TL;DR at the top of every response. Almost everything in lists or tables. Every section stands alone.
This is what the same Claude response looks like without and with hyperfocus.
Without hyperfocus (default Claude):
The
useMemohook in React is used to memoize expensive computations so that they are only recalculated when their dependencies change. This is particularly useful when you have a component that re-renders frequently but has some computation that doesn't need to run on every render. WithoutuseMemo, React would recalculate the value on every render, which could lead to performance issues. You should use it when you have computationally expensive operations that depend on specific props or state values, but be careful not to overuse it as the memoization itself has a small cost.
With hyperfocus — all three modes side by side:
| Clean | Flow | Zen |
|---|---|---|
|
Use it when the computation is expensive and the component re-renders frequently. Don't overuse it — memoization itself has a small cost. |
What It Does It caches a computed value. React only recalculates when dependencies change. When To Use It Use
Watch Out Don't overuse it. Memoization has a small cost. Only apply it where you can measure a difference. |
TL;DR: What
When
Caution
|
Claude Code marketplace (recommended):
claude plugin marketplace add nextor2k/hyperfocus
claude plugin install hyperfocus@hyperfocusWhen installed from the marketplace, the command appears as /hyperfocus:hyperfocus in the marketplace UI, but bare /hyperfocus usually works in your prompt.
Or install the skill directly:
npx skills add nextor2k/hyperfocusOther agents (Codex, GitHub Copilot, manual)
Codex:
npx skills add nextor2k/hyperfocus -a codexGitHub Copilot:
npx skills add nextor2k/hyperfocus -a github-copilotManual: Clone this repo. Copy skills/hyperfocus/ into your project's skills folder.
| Command | Effect | Scope |
|---|---|---|
/hyperfocus |
Activate in flow mode | This session only |
/hyperfocus clean|flow|zen |
Switch mode | This session only |
/hyperfocus off or stop hyperfocus |
Deactivate | This session only |
/hyperfocus persistent [mode] |
Enable global auto-activation | All future sessions |
/hyperfocus persistent project [mode] |
Enable project-scoped auto-activation | This project, all sessions |
/hyperfocus disable |
Disable all auto-activation | Both scopes (if they exist) |
/hyperfocus status |
Check session mode + persistent state | — |
You can also say "focus mode", "adhd mode", or "adhd friendly" in your prompt to trigger activation.
Want hyperfocus on every session without typing /hyperfocus each time?
/hyperfocus persistent
This enables a plugin-owned SessionStart hook that re-injects the formatting rules at the start of every session. Your settings.json is never modified — hyperfocus stores its state in a dedicated plugin data directory.
| Command | What it does |
|---|---|
/hyperfocus persistent |
Auto-activate globally (all projects) |
/hyperfocus persistent project |
Auto-activate for this project only |
/hyperfocus disable |
Remove auto-activation |
How it works: When you install hyperfocus from the Claude Code marketplace, the plugin ships a SessionStart hook that reads your persistent state at session start and re-injects the rules automatically. Your state lives in the plugin's data directory (separate from your settings). Uninstalling the plugin removes its state entirely.
Note: Persistent mode requires a POSIX shell (sh) and is available on macOS, Linux, and WSL. Standalone skill installations (via npx skills) do not include persistent hooks.
Hyperfocus is a skill — formatting rules that load into Claude's context. No code, no build step. Code blocks, commits, and PRs stay untouched. Only prose gets restructured.
| Feature | Marketplace Plugin | Standalone Skill |
|---|---|---|
| Session modes (clean/flow/zen) | ✓ | ✓ |
| Persistent auto-activation | ✓ | — |
| Project-scoped persistence | ✓ | — |
| SessionStart hook | ✓ | — |
If you're upgrading from an older version, you may need to clean up legacy configuration:
-
Remove old settings hook: If your
~/.claude/settings.jsoncontains aUserPromptSubmithook for hyperfocus-rules, delete it. The new version uses SessionStart instead. -
Remove old rules file: If
~/.claude/hyperfocus-rules.txtexists, you can safely delete it. Rules are now managed by the plugin.
After upgrading, persistent mode will work automatically with the new hook. No manual configuration needed.
- Claude Code — fully tested, mode persists all session
- Codex — supported, mode persists all session
- GitHub Copilot — supported via npx skills
Warning
Some agents (Opencode, omp) may need re-activation each turn. That's an agent limitation, not a hyperfocus one.
Hyperfocus is informed by cognitive-accessibility research and guidance.
| Principle | What the research says | Source |
|---|---|---|
| Short sentences | Comprehension drops sharply past 20 words | Oxford/Plain English |
| Chunking | ADHD brains process sequential chunks better than parallel info | PMC6996017 |
| Front-loading | ADHD readers may not reach the end — answer goes first | W3C COGA |
| Visual hierarchy | Good subheadings enable re-entry in under 2 seconds | ATG Publishing |
| Blank lines | Walls of text trigger task abandonment | ATG Level 1 |
Full source list
- W3C COGA — Making Content Usable for People with Cognitive and Learning Disabilities (2021)
- ATG Publishing — ADHD Friendly Publishing Standards
- WCAG 2.2 — Cognitive Accessibility (2023)
- MDN — Creating Effective Technical Documentation
- Readability Guidelines — Simple Sentences (Oxford Guide to Plain English)
- PMC — Constraints on Information Processing Capacity in Adults with ADHD (2020)
ADHD isn't a deficit to work around. Hyperfocus is the superpower.
This plugin removes the friction that blocks it. When information is structured right, your brain locks in — not because it was forced to, but because nothing got in the way.
If you have ideas for better ADHD-friendly formatting — especially if you have ADHD yourself — open an issue or PR.