What is the main benefit of using the authoring-prompts skill?

It provides a rigorous framework to write instructions that are executed more reliably and use fewer tokens, making your AI agents faster and more predictable.

When should I invoke this skill?

You should invoke this skill first whenever you are beginning to author new AI agent instructions, custom skills, or repository-wide rules.

Can I use these principles for my project's CLAUDE.md file?

Yes, the principles of determinism, specificity, and imperative language are specifically designed to optimize CLAUDE.md files for better agent interactions.

How does this skill help with token costs?

It applies principles of 'Token Economics' to identify and remove decorative language and filler text, ensuring you only send high-value context to the LLM.

What are instruction anti-patterns?

Anti-patterns are common authoring mistakes like using vague quantifiers (e.g., 'usually'), decorative politeness, or burying critical constraints in long paragraphs.

AI Instruction Authoring Foundations

Name: AI Instruction Authoring Foundations
Author: jkappers

byjkappers

生産性とワークフロー

Establishes a standardized framework for writing efficient, deterministic, and token-optimized instructions for AI agents and Claude Code skills.

概要

This skill serves as the foundational guide for developers creating custom AI instructions, CLAUDE.md files, and automated rules. By focusing on token economics, imperative language patterns, and structural optimization, it ensures that LLMs execute tasks with high determinism and minimal ambiguity. It helps users eliminate decorative language, specify precise command flags, and adopt formatting patterns that enhance machine parsing, ultimately leading to more reliable, predictable, and cost-effective AI agent interactions within the Claude Code environment.

主な機能

Guidelines for maintaining terminology consistency and constraint hierarchy.
Structural formatting templates using Markdown, tables, and lists for better parsing.
0 GitHub stars
Token-saving strategies to minimize context window usage and reduce filler.
Imperative language patterns for direct, deterministic agent behavior.
Identification and correction of common instruction anti-patterns and vague language.

ユースケース

Developing custom Claude Code skills with precise execution constraints and clear logic.
Refining a project's CLAUDE.md file to improve agent alignment and task performance.
Standardizing internal team documentation and rules for AI-assisted development workflows.

概要

主な機能

Guidelines for maintaining terminology consistency and constraint hierarchy.
Structural formatting templates using Markdown, tables, and lists for better parsing.
0 GitHub stars
Token-saving strategies to minimize context window usage and reduce filler.
Imperative language patterns for direct, deterministic agent behavior.
Identification and correction of common instruction anti-patterns and vague language.

ユースケース

Developing custom Claude Code skills with precise execution constraints and clear logic.
Refining a project's CLAUDE.md file to improve agent alignment and task performance.
Standardizing internal team documentation and rules for AI-assisted development workflows.