focus

A Claude skill for rigorous, detail-preserving summarization of academic papers and long-form documents.

Adapted from the FOCUS workflow described in Lin (2025), Nature Biotechnology. The method extracts every key point from a document, section by section, with specific details and naturally embedded direct quotes, then produces a clean final output with an overview at the top.

What it does

Most AI-generated summaries lose the details that make a paper worth reading: effect sizes, method names, exact comparisons, the author’s own language where it matters. This skill takes the opposite approach. It treats exhaustiveness as a feature, not a bug, and preserves the specificity of the source material.

The two-step process:

1. Extract every key point from the document, section by section, with concrete details and direct quotes embedded naturally (not restated alongside paraphrase).
2. Clean up by stripping citation markers, adding a concise overview/takeaway, and organizing sections for readability.

The output is a numbered list organized under section headings, with bold key terms and italicized quotes. No meta-commentary, no attribution to “the authors,” no filler.

Installation

Option 1: Download ZIP

Click the green Code button at the top of this repo, then Download ZIP. Extract the ZIP and add the folder to your Claude skills directory.

Option 2: Releases

Go to the Releases page and download the latest .skill file. Add it to your Claude skills in customize/skills on the web, or double-click it if you have Claude Desktop installed.

Option 3: Build it yourself

Build a .skill file from the source code, then add it to your Claude skills as described above.

git clone https://github.com/stephenturner/skill-focus.git
cd skill-focus
zip -r focus.skill SKILL.md references/

Usage

Trigger the skill with any of:

• /focus (followed by or alongside an uploaded document)
• “Summarize with FOCUS”
• “FOCUS summary”
• “Use the focus method”

It also triggers on requests for exhaustive or detail-preserving summaries where the user wants every key insight captured with supporting quotes.

Example output format

## Overview

[3-6 sentence paragraph summarizing the document's main contribution and findings.]

## Section Title

1. **Key finding in sentence case**
The experiment showed a 23% reduction in error rate across all conditions,
with the strongest effect in the high-complexity group. The method
*"outperformed all baselines on both precision and recall"* when applied
to datasets with more than 10,000 samples.

2. **Another insight**
...

What this skill is not