llm-docs-optimizer_skill

Overview

This skill optimizes project documentation and READMEs specifically for AI coding assistants and LLMs like Claude and GitHub Copilot. It improves documentation quality using c7score-guided transformations, generates llms.txt navigation files, restructures content around developer questions, and returns automated quality scores. Use it to make docs copy-paste runnable, LLM-friendly, and focused on real developer workflows.

How this skill works

The skill analyzes existing documentation and produces a prioritized plan: generate developer questions, map those questions to code snippets, and identify coverage gaps. It applies targeted optimizations (complete examples, consolidated snippets, correct language tags, removal of irrelevant metadata) and validates each snippet for runnable quality. For c7score optimizations it calculates before/after metric scores; when requested it also creates an llms.txt file that summarizes project structure for LLM navigation.

When to use it

• You want to optimize README/docs for c7score or Context7 benchmarking.
• You need an llms.txt navigation file so LLMs can find relevant docs quickly.
• You’re preparing documentation for LLM-assisted coding assistants (Claude, Copilot, etc.).
• You want automated quality analysis and actionable fixes for code snippets.
• You need README improvements to make examples copy-paste runnable and self-contained.

Best practices

• Prioritize question-driven examples: lead with "How do I..." and provide runnable snippets.
• Keep one lesson per snippet; combine tiny fragments into full workflows.
• Always include necessary imports and setup so examples are copy-paste ready.
• Use correct language tags and TITLE / DESCRIPTION / CODE structure for each snippet.
• Remove noise: licensing blocks, directory trees, and citation entries from main docs.

Example use cases

• Transform an API reference into a Quick Start with runnable Python examples.
• Audit a multi-file docs/ directory to find missing answers for common developer questions.
• Generate an llms.txt for a Python library so LLMs can prioritize Quick Start and Examples.
• Consolidate many small snippets into one comprehensive workflow for a CLI tool.
• Produce a c7score before/after report showing metric improvements after edits.