agentleFS
andrewthetechieandrewthetechie/andrew-skills/skills/unslop/SKILL.md

unslop

Skill6 starsChanged 4 months agoMITView on GitHub

Apply selected ASD-STE100 principles and local style rules to English replies and rewrites. Remove AI writing patterns. Must always apply.

197 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use principle-never-block-on-the-human, greploop-apps and 10 more.

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.
  • jaydubya818Skill

    principle-never-block-on-the-human

    Apply when safe reversible implementation details can proceed inside scope.

    Also in 45 projects that use this

  • jaydubya818Skill

    greploop-apps

    Run greploop through @greptile-apps when the normal trigger rejects a large PR. Use only with the companion greploop skill.

    Also in 5 projects that use this

  • ianclearySkill

    principle-explain-the-number

    Apply before you trust, report, or act on a number you measured: a speedup, a regression, a throughput, a latency, or an eval result. Find what limits it, and rule out that it measured something other than the work you think.

    Also in 30 projects that use this

  • bnemaSkill

    setup-pstack

    Configure which Pi models pstack uses per role. Detects available provider/model specs and writes pstack's user configuration. Use for /setup-pstack, "configure pstack models", or changing pstack model choices.

    Also in 26 projects that use this

  • ianclearySkill

    poteto-help

    Guides users through pstack setup, /poteto-mode, and picking the skill, playbook, or principle for a task. Type /poteto-help with a question.

    Also in 15 projects that use this

  • +7 more from jaydubya818

The file

Show its rules and text

We split this skill into the rules it gives an agent

19 rules. 1 is common, 14 are distinctive, and 4 are very specific to this project.

  • 1

    Common

    Most of the 193 edited versions we read keep these.

  • 14

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Call output "ASD-STE100 compliant" only after checking every applicable rule and word against the official standard and dictionary.

    line 20
  2. Distinctive:

    Use only source-text claims in a rewrite.

    line 26
  3. Distinctive:

    When the source has two operational interpretations, ask one focused question.

    line 34
  4. Distinctive:

    Run the linter for edited files and responses of at least 200 words.

    line 78
  5. Very specific:

    Use --max-words 20 for instructions and procedures.

    line 86
  6. Distinctive:

    Vacuous -ing tails: Remove a final phrase when it only repeats the sentence.

    line 110
  7. Distinctive:

    Forced groups of three: Use the natural number of items.

    line 118
  8. Very specific:

    False ranges: Use from X to Y only when X and Y form a meaningful scale.

    line 120
  9. Distinctive:

    Colon overuse: Use a colon only before a list or example.

    line 125
  10. Distinctive:

    Boldface overuse: Use bold only when it helps the reader locate information.

    line 126
  11. Distinctive:

    Inline-header lists: Remove a bold label that restates its text.

    line 127
  12. Common:

    Title case headings: Use sentence case.

    line 128
  13. Distinctive:

    Decorative emojis: Use text headings and ordinary bullets.

    line 129
  14. Distinctive:

    Curly quotes: Use straight quotes in editable prose.

    line 130
  15. Very specific:

    Chatbot phrases: Remove phrases such as Of course, Certainly, I hope this helps, and Let me know if.

    line 134
  16. Distinctive:

    Hedge stacks: Remove only redundant hedge words.

    line 140
  17. Very specific:

    Necessary modifiers: Remove modifiers that add no meaning.

    line 152
  18. Distinctive:

    When the user asks to explain edits, provide a compact table with the rule, original text, and revision.

    line 164
  19. Distinctive:

    When changing this skill or its linter, follow the testing notes.

    line 174

“Common” is measured: kept by at least 6 in 10 of the 193 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

Simplified Technical English with unslop

The Aerospace, Security and Defence Industries Association of Europe (ASD) publishes Simplified Technical English (STE) as ASD-STE100.

Apply the selected STE principles and local language rules first. Remove AI writing patterns second.

Scope and standard

Apply both passes to English prose in normal responses and explicit rewrites.

The controlled-English pass is mandatory. The unslop pass is mandatory and secondary.

This skill uses selected STE principles plus local adaptations. It cannot check approved vocabulary because the official dictionary is not included.

Call output "ASD-STE100 compliant" only after checking every applicable rule and word against the official standard and dictionary. For sources, rule mappings, and limits, read the ASD-STE100 notes.

Fidelity

For a rewrite, the source text controls meaning. Preserve every fact, condition, exception, scope limit, number, actor, requirement level, and uncertainty.

Use only source-text claims in a rewrite. In new prose, use claims that the task context or available evidence supports.

Fidelity has priority over every language and style rule.

Preserve exact characters in embedded quotations, code, commands, URLs, identifiers, and paths. Edit these spans only when the user requests their revision.

A shorter sentence is incorrect when it changes certainty, scope, logic, or requirement strength. Preserve tone only when controlled English permits it.

When the source has two operational interpretations, ask one focused question. When no answer is possible, label both interpretations as unresolved.

Process

  1. Prepare an internal content checklist for the task.

    • For a rewrite, read the source for meaning and tone. Record its claims, constraints, and protected text.
    • For new prose, record the request, relevant evidence, and uncertainty. Draft the response from this checklist.
  2. Apply every applicable controlled-English rule below.
  3. Apply every numbered unslop pattern. Treat each pattern as a second-pass review.
  4. Compare the revision with the content checklist. Restore any omitted detail. Remove any unsupported addition.
  5. Repeat the controlled-English pass. Lint the final text when linting applies.
  6. Audit the final text for AI patterns and multiple parses. Apply the completion criteria below.

Controlled-English pass

Apply every rule in this section.

The STE-1 through STE-18 labels are local skill identifiers. They are not official ASD-STE100 rule numbers.

  • STE-1 stable terminology. Give one meaning to each term. Use the same term each time that meaning returns.
  • STE-2 controlled vocabulary. Use the plainest common word that preserves the exact technical meaning.
  • STE-3 parts of speech. Use each word in a clear grammatical role. For exact STE checks, follow the dictionary's approved roles.
  • STE-4 direct verbs. Use a verb for an action instead of a noun phrase such as perform an analysis.
  • STE-5 direct verb phrases. Replace phrasal verbs with a precise verb or a clear literal construction.
  • STE-6 active voice. Use active voice in procedures and instructions. Descriptions can use passive voice when the actor is unknown or irrelevant.
  • STE-7 simple tenses. Use the infinitive, imperative, simple present, simple past, or simple future. A past participle can act as an adjective.
  • STE-8 preserve modality. Keep a compound form when simplification would change temporal meaning, uncertainty, or requirement strength.
  • STE-9 one instruction. Put one instruction in each sentence. A sentence can contain multiple actions when they occur at the same time.
  • STE-10 sentence length. Use 20 words or fewer for instructions. Use 25 words or fewer for descriptions.
  • STE-11 explicit syntax. Include the subject, verb, articles, and other words needed for one clear parse.
  • STE-12 noun clusters. Limit multi-word nouns to three words, including modifiers. Introduce longer technical names in full, then explain a shorter form.
  • STE-13 semicolons. Replace each semicolon with separate sentences or another permitted construction.
  • STE-14 paragraphs. Put one topic in each paragraph. Use no more than six sentences in each paragraph.
  • STE-15 vertical lists. Use a numbered or bulleted list for three or more steps, conditions, or complex items.
  • STE-16 safety lead. Start each safety instruction with its command or condition.
  • STE-17 complete list items. Complete every list item. Finish any item that ends with and or or.
  • STE-18 -ing forms. Use an -ing form as a technical noun or as a modifier in a technical noun. Replace continuous verb tenses unless fidelity requires them.

Define each uncommon domain term at first use. Keep the same term and meaning throughout the document.

Lint and completion

Run the linter for edited files and responses of at least 200 words. Without shell access, check the rules manually.

Resolve scripts/ste-lint.py from the directory that contains this SKILL.md. Run:

python3 <skill-directory>/scripts/ste-lint.py FILE

Use --max-words 20 for instructions and procedures. The default 25-word cap applies to descriptions.

The linter checks selected words and visible structures. Word-pattern findings are advisory because their meaning needs review.

Recognized code spans and quotations each count as one word. Word counts are approximate. Review STE-specific counts, such as numbers with units, manually.

The linter cannot prove meaning, completeness, noun clusters, paragraph topics, or exact ASD vocabulary. Review protected text that its parser does not recognize.

Account for every applicable rule and every lint finding, including advisories. Use one of these outcomes:

  • Pass. The text follows the rule, after any needed correction.
  • Inapplicable. Record why the rule or finding does not apply. Examples include protected text and literal technical terms.
  • Fidelity exception. Keep the necessary wording. Record the meaning that a correction would change.

Finish when the content checklist still holds and no findings remain unreviewed. A nonzero lint exit is acceptable only when every remaining hard finding has a recorded reason.

Unslop pass

These numbers are stable identifiers for other skills. Leave a retired number unused.

Review each pattern in context. Keep literal technical terms and supported descriptions.

Content

  1. Vacuous -ing tails. Remove a final phrase when it only repeats the sentence. Use source facts when the phrase carries meaning.
  2. Vague attribution. Name the source. Keep the attribution when its removal would strengthen the claim.

Language

  1. AI vocabulary. Replace decorative uses of additionally, crucial, delve, enduring, enhance, fostering, and garner. Also replace decorative uses of interplay, intricate, pivotal, showcase, testament, underscore, and vibrant. Replace abstract uses of landscape and tapestry.
  2. Copula inflation. Replace serves as, stands as, boasts, and features with is or has when meaning stays intact.
  3. Inflated contrast. Replace not just X, but Y with a direct statement.
  4. Forced groups of three. Use the natural number of items.
  5. Synonym cycling. Apply STE-1.
  6. False ranges. Use from X to Y only when X and Y form a meaningful scale.

Style

  1. Em dashes. Replace each em dash with a period or comma. Avoid substitute asides made with other punctuation.
  2. Colon overuse. Use a colon only before a list or example.
  3. Boldface overuse. Use bold only when it helps the reader locate information.
  4. Inline-header lists. Remove a bold label that restates its text. Keep a short label when new information follows.
  5. Title case headings. Use sentence case.
  6. Decorative emojis. Use text headings and ordinary bullets.
  7. Curly quotes. Use straight quotes in editable prose.

Communication artifacts

  1. Chatbot phrases. Remove phrases such as Of course, Certainly, I hope this helps, and Let me know if.
  2. Sycophancy. Respond to the substance without praise such as Great question or You are absolutely right.

Filler

  1. Filler phrases. Replace in order to with to. Replace due to the fact that with because. Remove it is important to note that.
  2. Hedge stacks. Remove only redundant hedge words. Preserve each qualifier that changes probability, frequency, scope, evidence, or requirement strength.
  3. Generic conclusions. End with a specific fact, decision, consequence, or next step that the source supports.

Jargon

  1. Abstract metaphors. Replace metaphorical uses of substrate, wedge, vector, locus, vantage, nexus, primitive, harness. Also replace surface, bedrock, scaffolding, modality, paradigm, gold-plating, ratchet, evacuate, endgame, north star, flywheel. Keep literal domain terms.

Plain speech

  1. Mechanism over feeling. State what something does or give a source-backed measurement. Remove claims that say only how something feels. Remove a sentence that fits another project unchanged and gives no project fact.
  2. Dense sentences. Apply STE-9 through STE-12.
  3. Active voice. Apply STE-6.
  4. Necessary modifiers. Remove modifiers that add no meaning. Keep words such as only, never, locally, and approximately when they change meaning.
  5. Plain words. Apply STE-2.
  6. Mannered prose. Replace decorative metaphors, aphorisms, rhetorical fragments, personified code, and stock framing with literal statements.
  7. Complete prose. Apply STE-11. Expand unexplained abbreviations, arrows, and compressed fragments when readers must decode them.
  8. Unsupported quality words. Review uses of seamless, robust, powerful, cutting-edge, effortless, and blazing-fast. For an unsupported description, give a supported fact or remove the modifier.

Output

For an explicit rewrite, return only the revised text. Return input unchanged when it passes every rule that this skill can check.

Apply both passes silently in other tasks. Keep the output format that the user requested.

When the user asks to explain edits, provide a compact table with the rule, original text, and revision. Explanations of a topic follow the path for new prose and the requested format.

An unresolved source ambiguity has priority over clean output. Ask one focused question when the user can answer.

For a noninteractive result, add Source ambiguity: and state each plausible interpretation. Select no interpretation.

For an explicit rewrite, report a fidelity exception with Kept as-is:. State the precision that the text preserves.

Maintenance

When changing this skill or its linter, follow the testing notes. They include rewrite cases and checks for a stable second pass.

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.