agentleFS
toonoobytoonooby/ste100-plugin/skills/ste100/SKILL.md

ste100

Skill0 starsChanged 41 days agoMITView on GitHub

Reference for writing or checking text in ASD-STE100 Simplified Technical English at a 10-100% compliance level. Use when the user asks for STE, Simplified Technical English, STE100, simpler technical prose, or a compliance check or rewrite of documentation, procedures, or replies. Also use when the user sends a level message such as "ste 70" or "/ste 70": reaching you means the plugin's hook did not run.

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 10 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Keep every caveat, uncertainty, edge case, contradiction, and unusual behavior.

    line 41
  2. Distinctive:

    If a rule forces a loss or change of meaning, break the rule.

    line 42

“Common” is measured: kept by at least 6 in 10 of the 10 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

ASD-STE100 writing at graded compliance

ASD-STE100 is a controlled language for technical documentation. It limits vocabulary and grammar so that readers (including non-native English speakers) understand text quickly and with one meaning. The official specification and dictionary are free from ASD (https://www.asd-ste100.org). This skill paraphrases the main writing rules; it does not contain the dictionary.

The slider

The plugin maps a level from 10% to 100% onto cumulative rule tiers. The user sets the level by sending ste <level> or /ste <level> as a message, and a hook adds the rules for that level to each turn.

If a level message reaches you

The hook normally catches valid level messages such as ste 70 and /ste off before they reach the model. If one reaches you, it was not intercepted. Do not set the level yourself; that would hide the problem. Check for a hook diagnostic and explain the required repair or approval.

  • First check for a hook diagnostic about a missing jq dependency or malformed input. Follow that diagnostic's repair instructions if it is present.
  • In Codex, plugin hooks stay off until the user trusts them. In the CLI, tell the user to run /hooks and trust the ste100 hook. In the desktop app, tell the user to review and trust it in Hooks settings. Then send the message again.
  • In Claude Code, the usual cause is that the plugin is disabled or the session started before it was installed. Tell the user to check /plugin and start a new session.
LevelAdds
10One idea per sentence; consistent terms; prefer active voice
20Sentences of 30 words or fewer
30Main point first; numbered lists for sequences
40Imperative instructions; 20-word procedural / 25-word descriptive limits; no phrasal verbs
50No idioms, metaphors, slang, or humor; keep articles and "that"; noun clusters of 3 words max; no -ing nouns/adjectives
60Paragraphs of 6 sentences max, one topic each; passive only in descriptive text when the agent is unknown or unimportant
70Only simple present, simple past, simple future, imperative, infinitive; no contractions; deviations must be marked with ‡
80Common concrete words, one meaning and one part of speech each; warnings before the step, starting with a command
90One instruction per sentence unless actions are simultaneous; topic sentence first in descriptive paragraphs
100Dictionary-approved words only, plus technical names and technical verbs of the field

The current hook block replaces all earlier automatic STE instructions. If its level is 0, stop applying those earlier rules and their deviations footer. An explicit user request to write or rewrite in STE still applies. At a lower active level, use only the current block's tiers.

The accuracy rule (overrides every tier)

STE can push a writer to smooth over complexity. That is a failure, not compliance.

  • Keep every caveat, uncertainty, edge case, contradiction, and unusual behavior.
  • If a rule forces a loss or change of meaning, break the rule.
  • Say that strange results are strange. Do not normalize them.
  • At 70% and above, mark each deviation with ‡ and list the reasons under "STE deviations:" at the end.

Out of scope

Code, commands, paths, identifiers, error messages, quoted text, tool inputs, and file contents stay as they are unless the user asks otherwise.

Rewrite examples

Original: "You'll want to go ahead and spin up the dev server before kicking off the tests, since they're hitting it."

  • 40%: "Start the dev server before you run the tests. The tests send requests to it."
  • 80%: "Start the development server. Then run the tests. The tests send requests to the server."

Original: "This mostly works, but weirdly the cache sometimes returns stale data after a deploy — we haven't pinned down why."

  • 70% (correct): "The cache usually returns correct data. After a deployment, the cache sometimes returns old data. This is unusual. We do not know the cause."
  • 70% (wrong — smooths over the problem): "The cache returns data after a deployment."

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.