wtf is simpleenglish?

SimpleEnglish is an agent skill that forces AI models to write documentation in ASD-STE100 Simplified Technical English, a controlled form of English that the aerospace industry has used since 1983 so a tired mechanic cannot misread an instruction. The project's pitch is that following these rules also removes most of the vague, padded, marketing sounding writing that AI models tend to produce by default. The README shows several before and after examples using real Claude output. A sentence that originally described a tool using fancy language to automatically move data between systems becomes a short, direct statement of what the tool does and what you need to set up first. An error message that apologizes for inconvenience and asks you to try again becomes a specific statement of what failed and how to fix it. The rules behind these rewrites include limits like twenty words per instruction, using only one word for one meaning throughout a document, sticking to simple verb tenses, avoiding words like should, would, may, and might, and putting a condition before the command it applies to rather than after. It works as an Agent Skill, a standard supported by many coding tools including Claude Code, Cursor, VS Code Copilot, OpenAI Codex, Gemini CLI, and roughly two dozen others. Installing it is a single command using the skills command line tool, or it can be added as a Claude Code plugin, or set as an always on output style inside Claude Code so all of its writing follows these rules. For tools that do not support the Agent Skills format, the repository includes a plain system prompt file that can be pasted into a system prompt, an AGENTS.md file, or custom instructions for tools like ChatGPT or Gemini. Beyond documentation, the skill also includes adapted rule sets for error messages, runbooks, and incident reports, each following a similar what happened, why, and what to do structure. The project is released under the MIT license and includes benchmark results claiming a large measured reduction in rule violations across six different Claude models.