AGENT SKILL

attention-span

ADHD-friendly output styles and attention-kind conciseness rules for AI agents (Claude Code, Codex, Antigravity, Cursor, Windsurf, Copilot). Provides Attention-kind, Spartan, and Rundown output styles for scannable, token-efficient agent communication.

License 许可
AGPL-3.0
最近更新
2026-11-14
Stars
★ 0
Forks
1

01 前言 Front Matter

name
attention-span
tools
Read, Write, Edit, Bash, Grep, Glob
license
AGPL-3.0
metadata
{ "origin": "alexgreensh/attention-span", "version": "0.6.0" }
description
ADHD-friendly output styles and attention-kind conciseness rules for AI agents (Claude Code, Codex, Antigravity, Cursor, Windsurf, Copilot). Provides Attention-kind, Spartan, and Rundown output styles for scannable, token-efficient agent communication.

02 资源文件 6

03 SKILL.md 正文

Attention Span Skill

A collection of ADHD-friendly output styles and conciseness guidelines for AI coding agents that optimize for reader scannability and attention retention without compromising engineering quality.

When to Activate

  • Asking an agent to format responses for maximum scannability and plain-English clarity
  • Configuring or switching agent output styles (Attention-kind, Spartan, Rundown)
  • Reducing token verbosity and eliminating AI fluff/preambles in agent responses
  • Setting up global or project-level output style rules across AI agents (Claude Code, Codex, Antigravity CLI, Cursor, Windsurf, Copilot)

Available Output Styles

Style Description Ideal For
Attention-kind ADHD-friendly, front-loaded answers, heavy bolding, arrow markers (→), spaced out blocks General pair programming, deep focus, preventing wall-of-text fatigue
Spartan Zero-warmth, blunt, imperative, maximum signal-to-noise ratio Heads-down work, rapid execution, quick technical answers
Rundown Briefing style with TL;DR, status checklist (✅/🟡/⬜), action items tagged with emoji Standups, status updates, progress reports

Core Guidelines & Principles

1. Attention-Kind Principles

  • Lead with the bottom line in line 1. The first sentence contains the direct answer.
  • Say the least that fully answers. Eliminate preambles ("Great question!", "Sure thing!"), restating questions, or closing restatements.
  • Build for scanning:
    • Mark key points with **→ Lead-in.** on a new line with blank lines between paragraphs.
    • Bold the main takeaway and key terms/numbers so skimming only the bold text yields the complete answer.
  • Preserve critical details: Never cut a warning, caveat, exact threshold, or scoped condition to save words.
  • Depth requests suspend brevity: If the user asks to "explain deeply", "walk me through it", or "give full context", brevity constraints are suspended for that turn.

2. Spartan Principles

  • Imperative & direct: Strip all warmth, transitions, and conversational fluff.
  • Deliverable purity: When asked for code, a commit message, or a document, return only the requested artifact without wrapping text.
  • One idea per paragraph block.

3. Rundown Principles

  • Open with TL;DR: carrying the complete summary.
  • State checklist: ☑️ Done, 🟡 In Progress, ⬜ Not Started.
  • Group next choices under Your move: with relevant emoji markers.
  • Never invent status: Report only verified facts.

Multi-Agent Configuration

Claude Code

Output styles can be selected via /style:

# Global output styles directory
mkdir -p ~/.claude/output-styles ~/.claude/commands
cp output-styles/*.md ~/.claude/output-styles/
cp commands/style.md ~/.claude/commands/

To set as default in ~/.claude/settings.json:

{
  "outputStyle": "Attention-kind"
}

Codex

Append the body of the desired output style to ~/.codex/AGENTS.md (idempotent block):

<!-- attention-span:start -->
[Attention-kind instructions body]
<!-- attention-span:end -->

Antigravity CLI (agy)

Append the style body to GEMINI.md in the current project or home directory:

<!-- attention-span:start -->
[Attention-kind instructions body]
<!-- attention-span:end -->

Windsurf / Devin / Cursor

For Windsurf/Devin global memories: ~/.codeium/windsurf/memories/attention-kind.md For Cursor global rules: ~/.cursor/rules/attention-kind.mdc or ~/.cursor/rules/


Repository & Links

04 同仓库更多 Skills

sanjeevafk
sanjeevafk/agent-skills

zarr-python

Chunked N-D arrays for cloud storage (Zarr-Python 3). Compressed arrays, parallel I/O, S3/GCS via fsspec, NumPy/Dask/Xarray compatible, for large-scale scientific computing pipelines.

★ 0
sanjeevafk
sanjeevafk/agent-skills

whiteboard-diagrams

Create accurate, simple, hand-drawn whiteboard diagrams from a user's process, system, architecture, learning workflow, comparison, or causal explanation. Use when the user asks for a scribbly whiteboard visual, educational workflow graphic, hand-drawn explainer, or diagram like the bundled reference; ask one focused clarification question when the subject or relationships are missing, then generate and visually verify the final image.

★ 0
sanjeevafk
sanjeevafk/agent-skills

webapp-testing

Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

★ 0
sanjeevafk
sanjeevafk/agent-skills

web-artifacts-builder

Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.

★ 0
sanjeevafk
sanjeevafk/agent-skills

verification-loop

A comprehensive verification system for Claude Code sessions.

★ 0
sanjeevafk
sanjeevafk/agent-skills

vercel-deploy

Deploy applications and websites to Vercel. Use when the user requests deployment actions like "deploy my app", "deploy and give me the link", "push this live", or "create a preview deployment".

★ 0