de-slopping

A finishing pass that strips AI-slop from writing before the reader sees it. Inspired by ASD-STE100 Simplified Technical English, the aerospace standard for flight manuals, where one misread sentence can kill someone.

AI prose fails the opposite way. Sentences run long. Claims hedge. Big words do small jobs. Filler sounds like content. This pass removes that. A linter catches the countable slop. I apply the rest by judgment.

When to run it

First classify the text. The mode decides how hard you cut.

• Strict (every rule): PR and commit messages, design docs, READMEs, runbooks, reports, release notes. Any artifact a reader acts on.
• Light (cut slop, hedges, em-dashes, and the swap words; keep voice, rhythm, and normal sentence length): chat replies, explanations, teaching.
• Exempt (do not touch): code, commands, error text, quoted text, file paths, URLs. Those are exact strings.

Full STE strips personality by design. Do not apply strict mode to prose that needs a voice. A de-slop pass with no light mode is a footgun.

The pass

1. Classify the text as strict, light, or exempt.
2. Run the linter on the draft: python3 ~/.claude/skills/de-slopping/ste_lint.py <file>, or pipe text on stdin.
3. Fix every hard finding. Judge each soft finding: keep it or cut it.
4. Apply the judgment rules below.
5. Re-run the linter. Confirm zero hard findings.
6. Send the clean text. Stay quiet about the edit unless asked for an audit.

With the other writing skills

For a deliverable, run the passes in this order:

1. minto-pyramid's buried-lede move, if the text argues, reports or asks for a decision (a PR description, a report, a proposal, an update): put the conclusion first. Keep every fact (move a detail lower instead of cutting it), and do not invent reasons to reach three. Skip this pass for reference docs, runbooks and narrative, as that skill says.
2. de-slopping (this skill): cut the slop and the hedges. Every fact stays.
3. humanizer: remove the AI tells that remain.
4. Run the linter again and fix every hard finding. The last two passes can bring back long sentences.

Rules the linter enforces (mechanical)

The linter counts words and matches patterns. It flags:

• Sentences over 25 words. Split them into numbered steps.
• The slop blocklist. Delete on sight:

  delve  crucial  seamless  comprehensive  robust  cutting-edge
  game-changer  tapestry  realm  plethora  myriad  elevate
  foster  streamline  meticulous  intricate  pivotal  testament

• One-word swaps:

  utilize, leverage        -> use      facilitate      -> help
  commence, initiate       -> start    terminate       -> stop
  prior to                 -> before   in order to     -> to
  additionally, moreover   -> also     approximately   -> about

• Stock hedges:

  it's important to note   it's worth mentioning
  when it comes to         at the end of the day

• Every em-dash. Use a comma, a colon, or a period.
• Soft modals (should, would, may, might): a soft flag. Keep can, will, must. They commit.
• Filler -ing such as making it easy to: a soft flag.

The linter matches exact forms. It flags commence but not commencing, and comprehensive but not comprehensively. I catch those variants by judgment.

Rules I apply (judgment)

• One word, one meaning. If a word reads two ways in context, replace it.
• One instruction per sentence. Sequences become numbered lists.
• Active voice. Name the agent: The pump moves the fuel.
• Condition before command. Write If the flag is set, press enter, not the reverse. The reader meets the condition first.
• Keep the small words. the, a, that stay. Short is not the same as broken.
• Preserve every fact. Add none. If a fact is missing, mark it [value missing].

What it refuses to do

• It does not delete facts to hit a word count. Forty facts in, forty out.
• It does not touch code, commands, or error strings. Those are exact.
• It does not invent a fact to finish a sentence. It marks the gap instead.

Modes

• Rewrite (default): output the clean text. Keep every fact.
• Audit: output only the findings. Give rule, quoted text, and fix, one line each. Change nothing.
• Light: apply the slop, hedge, em-dash, and swap rules only. Keep the voice. Say that you used light mode.

Example

Before (44 words, one sentence):

In order to facilitate the seamless integration of the authentication module, it is important to note that developers should ensure that the configuration file has been comprehensively updated prior to commencing the deployment process, which can be accomplished via the CLI.

After (longest sentence 9 words):

Update the configuration file before you start the deployment. Use the CLI to deploy. This connects the authentication module.

Same facts. A third of the words. No ambiguity.

Attribution

Inspired by ASD-STE100 Simplified Technical English, maintained by the AeroSpace and Defence Industries Association of Europe (ASD). This skill applies its principles. It is not the official specification or its dictionary. The specification is free at asd-ste100.org.