agentleFS
geekxxgeekxx/ai-assistant-starter-kit/.claude/skills/ste/SKILL.md

ste

Skill1 starsChanged 4 months agoMITView on GitHub

Rewrite INTERNAL procedural docs (runbooks, SOPs, onsite/setup checklists, step-by-step ops instructions) into ASD-STE100 Simplified Technical English — terse, one-instruction-per-sentence, deliberately flat. ONLY runs when the user explicitly invokes it by name — "/ste" or "use the ste skill" / "apply ASD-STE100". It is never auto-triggered on paraphrased intent ("simplify this," "make it clearer," "shorter sentences"). SCOPE BOUNDARY: use ONLY for internal procedural artifacts. Do NOT apply it to client- or firm-facing prose (emails, recaps that travel, Op Mech notes, briefs) or to anything that goes through the humanizer pass — STE and the humanizer are contradictory. If a doc needs the humanizer, it is not an STE document.

18 other projects use a version of this, 3 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use interview.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • dazuckSkill

    interview

    Interview me about a plan to flesh out requirements before implementation. Asks non-obvious questions one at a time until all ambiguities are resolved.

    Also in 3 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. 0 are common, 12 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 16 edited versions we read keep these.

  • 12

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this ONLY for internal procedural artifacts: runbooks, SOPs, onsite/setup checklists, step-by-step ops instructions — anywhere terse, …

    line 11
  2. Distinctive:

    If you are unsure whether a doc is procedural-internal or travels externally, ask Jeff before applying STE.

    line 13
  3. Distinctive:

    Before writing a single sentence, decide: is this procedural text (instructions someone follows) or descriptive text (explanation, …

    line 21
  4. Distinctive:

    Put a condition BEFORE its command: "If the pressure decreases, close the valve."

    line 30
  5. Distinctive:

    Do not omit articles, subjects, or verbs to save words.

    line 31
  6. Distinctive:

    Never use present perfect or continuous forms.

    line 36
  7. Distinctive:

    Never use an -ing form as a verb.

    line 37
  8. Distinctive:

    No phrasal verbs: "go down" → "decrease," "set up" → "install," "carry out" → "do."

    line 42
  9. Very specific:

    Before drafting, replace unapproved vocabulary.

    line 46
  10. Distinctive:

    No Latin abbreviations: "e.g." → "for example," "i.e." → "that is," delete "etc."

    line 50
  11. Distinctive:

    No semicolons — write two sentences.

    line 53
  12. Distinctive:

    Start a warning or caution with the command or condition, then give the risk: "WARNING: Do not touch the terminal.

    line 60
  13. Distinctive:

    After drafting, scan your text once for each of these and fix every hit before you respond:

    line 66

“Common” is measured: kept by at least 6 in 10 of the 16 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 Simplified Technical English

Scope and boundary — read first

  • Use this ONLY for internal procedural artifacts: runbooks, SOPs, onsite/setup checklists, step-by-step ops instructions — anywhere terse, unambiguous, one-instruction-per-sentence text is the goal.
  • Do NOT apply STE to client-facing or firm-facing prose (emails, recaps that travel, Op Mech notes, briefs), and do NOT combine it with the humanizer skill. The two are contradictory: STE is deliberately flat and repetitive, and the humanizer makes prose read naturally. If a document needs the humanizer, it is not an STE document.
  • If you are unsure whether a doc is procedural-internal or travels externally, ask Jeff before applying STE.

Apply the ASD-STE100 standard to all prose you produce in this task. Do not announce that you use STE, do not name the standard, and do not explain the style unless the user asks. If the user later asks you to "write more naturally," ask one short question to confirm they want to leave STE before you drop it.

Compliance note (for you, not for output): the official specification and its dictionary are copyright ASD. This skill encodes paraphrased rules and a publicly sourced word list. For certified aerospace/defense deliverables, tell the user that full compliance requires the free official specification (asd-ste100.org) and a human sign-off. Never claim certified compliance.

Step 0 — Classify the text

Before writing a single sentence, decide: is this procedural text (instructions someone follows) or descriptive text (explanation, background, description)? Every limit below depends on this. Mixed documents get classified section by section.

Core rules

Sentences

  • Procedural: maximum 20 words per sentence.
  • Descriptive: maximum 25 words per sentence.
  • Maximum 6 sentences per paragraph. One topic per paragraph.
  • One instruction per sentence. Two actions in one sentence only if they occur at the same time.
  • Put a condition BEFORE its command: "If the pressure decreases, close the valve."
  • Do not omit articles, subjects, or verbs to save words. "Ensure file exists" is wrong; "Make sure that the file exists" is correct. Keep the word "that" after verbs like "make sure."
  • Numbers, units with numbers, abbreviations, quoted strings, code identifiers, and proper nouns each count as one word.

Verbs

  • Allowed forms only: infinitive, imperative, simple present, simple past, simple future, and past participle used as an adjective.
  • Never use present perfect or continuous forms. "We have received" → "We received." "is being tested" → a simple form.
  • Never use an -ing form as a verb. An -ing word is allowed only inside a technical name ("the mounting bracket," "logging").
  • Active voice. Passive is allowed only in descriptive text when the agent is unknown or unimportant.
  • Instructions use the imperative: "Open the panel," not "You must open the panel" or "The panel should be opened."
  • Express actions as verbs, not nouns: "compress the file," not "perform compression of the file."
  • Modals: use can (possibility), will (future), must (requirement). Do not use should, would, could, may, might. A hedge becomes a fact or a "can": "an explosion can occur."
  • No phrasal verbs: "go down" → "decrease," "set up" → "install," "carry out" → "do."

Words

  • One word, one meaning, one part of speech, used consistently. Never rotate synonyms: pick one name for a thing and repeat it.
  • Before drafting, replace unapproved vocabulary. Read references/word-substitutions.md and apply it; it is the working dictionary for this skill.
  • Domain-specific nouns (part names, tool names, product names, UI labels) and domain verbs (drill, ream, boot, compile) are your technical nouns/verbs — keep them as-is, use each consistently, and do not verb a noun or noun a verb.
  • Noun clusters: maximum 3 words ("overhead panel light" is the limit). Longer clusters get decomposed with prepositions or hyphenated on first use: "main-gear-door retraction-winch handle."
  • American English spelling.
  • No Latin abbreviations: "e.g." → "for example," "i.e." → "that is," delete "etc."

Punctuation

  • No semicolons — write two sentences.
  • Parentheses only for references, abbreviations, and item numbers.
  • Hyphenate words that act as one unit; a hyphenated word counts as one word.
  • No contractions.

Warnings, cautions, notes

  • WARNING = risk of injury or death. CAUTION = risk of damage. NOTE = information only, never an instruction.
  • Start a warning or caution with the command or condition, then give the risk: "WARNING: Do not touch the terminal. The terminal has a dangerous voltage."
  • Notes obey the 25-word descriptive limit.

Step 2 — Self-check pass

After drafting, scan your text once for each of these and fix every hit before you respond:

  1. Any sentence over the 20/25-word limit for its type
  2. Contractions, semicolons
  3. "should," "would," "could," "may," "might"
  4. "has been," "have been," "had been," "is being," "was being"
  5. -ing words used as verbs
  6. Missing articles (a/an/the/this) before nouns
  7. Synonym rotation (the same object under two names)
  8. Any word in the unapproved column of references/word-substitutions.md
  9. Warnings that state the risk before the command

Reference files

  • references/word-substitutions.md — unapproved → approved word mappings and one-meaning rulings. Read it before drafting; it is short.
  • references/examples.md — worked before/after rewrites (procedural, descriptive, warnings, common mistakes). Read it when rewriting existing text or when unsure how a rule applies.

What NOT to touch

Code blocks, command strings, file paths, error messages, quoted UI text, and proper nouns stay exactly as written. STE applies to the prose around them.

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.