danyuchn/asd-ste100-skill

What it solves

This project addresses the problem of ambiguity in English text that is parsed by AI agents. When one agent's output is used as input for another agent, dense or vague language can lead to misinterpretation because agents lack the a human's ability to ask for clarification. It applies the principles of ASD-STE100 Simplified Technical English (STE)—a standard originally designed for aerospace maintenance—to ensure instructions and descriptions are literal, flat, and unambiguous.

How it works

The tool is a skill for Claude Code that rewrites text based on two modes: Strict (for procedures and error messages) and STE-flavored (for READMEs and prose). It analyzes input text sentence-by-sentence to flag violations such as passive voice, complex tenses, ambiguous word choices, and oversized noun clusters. It then rewrites these sentences to be clear and direct without losing factual precision. While it does not redistribute the official ASD dictionary, it applies the underlying principles of using the plainest available word consistently.

Who it’s for

It is designed for developers and AI architects creating agent-to-agent communications, tool/function descriptions, error messages, and system prompts where machine-readability and precision are critical.

Highlights

• Two-mode rewriting: Offers a strict mode for technical procedures and a more flexible flavored mode for general prose.
• Criterial analysis: Flags specific mechanical rule violations like phrasal verbs, semicolons, and hedge stacks.
• Agent-centric design: Specifically optimized for inter-agent instructions to prevent parsing errors.
• Precision-first: Prioritizes factual accuracy over brevity, flagging trade-offs rather than silently simplifying content.