Back to directory

attention-span

A small set of alternative output styles for Claude Code that change how it talks, not how it codes, trimming responses to plain, answer-first language for readers who find long output hard to follow.

attention-span

attention-span is a small collection of output styles for Claude Code that change how it talks to the user rather than how it codes; each is a single markdown file that is dropped in and switched on. Three styles are included: Attention-kind, an ADHD-friendly style that answers first, keeps responses short, drops jargon, and marks the important words in bold so a reader can skim just the bold text; Spartan, a terse style with no added warmth; and Rundown, a TL;DR-style briefing format.

The project frames shorter output as a side effect of being kind to a reader's attention, rather than the primary goal, and gives a worked before-and-after example showing a roughly 430-word default answer condensed to about 94 words in the Attention-kind style with the same underlying engineering content.

It is written in Python, documented in English, Spanish and Chinese, and released under the AGPL-3.0 license.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.