agentleFS
momentmakermomentmaker/kaijutsu/skills/core/deslop/SKILL.md

deslop

Skill3 starsChanged 6 months agoMITView on GitHub

Strip AI-generated tells from a draft without changing its substance. Use when the user says "deslop", "remove AI tells", "make this sound less like ChatGPT", or invokes /deslop. Composed by polish, journal, readme-update, and any skill that produces user-facing prose.

118 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use simple-issue-description, control-cli 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.
  • JwuthriSkill

    simple-issue-description

    Turn a rough bug report, feature request, support note, or pull request into a short, plain-language issue focused on the problem and desired behavior. Use when a contributor asks to simplify an issue, explain what a PR is for, create the corresponding issue for a PR, remove implementation detail from a report, or invokes /simple-issue-description.

    Also in 4 projects that use this

  • ElyzeSolutionsSkill

    control-cli

    Build or adapt a local harness to drive, inspect, and profile an interactive CLI or TUI without external services. Use for CLI UX checks, startup regressions, memory leaks, hangs, prompt flows, or terminal demos.

    Also in 6 projects that use this

  • JwuthriSkill

    openseo-review-web-content

    Write and review content for the OpenSEO website (web/) — blog posts, guides, feature pages, FAQs. Distills the philosophy for on-brand, useful, accurate content. Use whenever adding or editing user-facing prose in web/content or web/src.

    Also in 4 projects that use this

  • learnerLjSkill

    verify-this

    Verify a claim with fresh local evidence: restate it falsifiably, capture baseline and treatment, compare artifacts, and return VERIFIED, NOT VERIFIED, or INCONCLUSIVE.

    Also in 6 projects that use this

  • ElyzeSolutionsSkill

    control-ui

    Build or adapt a local browser/CDP harness to drive and inspect a web, IDE, or Electron UI. Use for local UI verification, screenshots, accessibility snapshots, perf profiles, visual diffs, or reproducing UI bugs.

    Also in 7 projects that use this

  • ten9876Skill

    papercuts

    Read-only AetherSDR issue triage that surfaces the UX papercuts, bugs, and regressions you can fix AND prove today with the agent automation bridge. Use when the user says $papercuts or /papercuts, or asks to "find papercuts", "what can I fix and prove today", "triage bridge-provable issues", "what UX bugs can the bridge close", "sort issues by priority I can knock out", or otherwise wants a curated, lifecycle-aware (design→code→prove) shortlist of issues — NOT the label-and-respond /triage workflow. This is the discovery step that feeds /fb. Strictly read-only: it never labels, comments, closes, or edits issues.

    Also in 4 projects that use this

  • +5 more from Jwuthri

The file

We split this skill into the rules it gives an agent

10 rules. 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 116 edited versions we read keep these.

  • 9

    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:

    Don't remove every emdash.

    line 28
  2. Distinctive:

    If a paragraph is being converted to a bullet list with 6+ items where 2 sentences would do, revert.

    line 54
  3. Distinctive:

    Domain-specific terms: If the draft uses "tarball," keep it.

    line 67
  4. Distinctive:

    The author's voice: If the original has a quirk that's intentional (a recurring metaphor, a specific tone), preserve it.

    line 68
  5. Distinctive:

    Markdown structure: when load-bearing — code blocks, headings, links.

    line 69
  6. Distinctive:

    Don't edit. Scan the draft and report flagged passages with category + suggested replacement.

    line 77
  7. Distinctive:

    Before applying any edits, read the project's CLAUDE.md / AGENTS.md / contributor docs for project-specific style guidance.

    line 80
  8. Very specific:

    If the project has a .deslop.yaml (or similar) configuration file at the repo root, that takes precedence over generic rules.

    line 86
  9. Distinctive:

    If you stripped something the user might want back, mark it: "I removed X — restore if intentional."

    line 94
  10. Distinctive:

    Don't add new content while deslopping: No new examples, no new caveats.

    line 98

“Common” is measured: kept by at least 6 in 10 of the 116 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

deslop

Models leak stylistic tells: emdash overuse, formulaic contrasts ("It's not X, it's Y"), pseudo-profound openers ("At its core..."), bullet-list explosions, and hollow connector words. Deslop removes those without changing what the text says.

When to invoke

  • Another skill produces prose for human consumption (commit messages, READMEs, journal entries)
  • The user explicitly asks: "deslop this", "remove AI tells", "make it sound less like ChatGPT"
  • A draft is going public (PR description, blog, docs)

What to remove

1. Emdash overuse

Models use emdashes (—) as a default structural break. Replace with periods, parentheses, or semicolons depending on weight.

OriginalDeslopped
"It's a great tool — fast, lightweight, and cross-platform.""It's a great tool. Fast, lightweight, cross-platform."
"We considered Postgres — but ultimately chose SQLite.""We considered Postgres but ultimately chose SQLite."
"The key insight — that caching invalidates predictably — drove the design.""The key insight (caching invalidates predictably) drove the design."

Don't remove every emdash. Keep the rare one where it carries real weight.

2. Formulaic contrasts

The "It's not X, it's Y" pattern. Or "Not just X but also Y." Replace with the direct positive claim.

OriginalDeslopped
"It's not just a CLI; it's a workflow.""It's a workflow CLI."
"Not merely fast — it's instant.""It's instant."

3. Pseudo-profound openers

"At its core...", "Fundamentally...", "In essence...", "Simply put...", "The thing about X is...". Cut and start with the actual claim.

OriginalDeslopped
"At its core, kaijutsu is a registry.""kaijutsu is a registry."
"Fundamentally, this changes how we think about agents.""This changes how we think about agents."

4. Hollow connectors

"It's worth noting that...", "It should be mentioned that...", "It's important to recognize...". If the next sentence is worth saying, just say it.

5. Bullet-list explosion

If a paragraph is being converted to a bullet list with 6+ items where 2 sentences would do, revert. Lists should reduce cognitive load, not perform structure.

6. Trailing summaries that restate the obvious

"In summary, kaijutsu provides...", "Overall, this approach...". Drop unless the doc is over 1000 words and a real summary would help.

7. The phrase "kaijutsu's whole pitch", "the whole game", "the magic is", and other gestures-at-importance

Just say what's important.

What to keep

  • The substance. Deslop changes voice, not claims. If the original says "we use SQLite," the deslopped version still says "we use SQLite."
  • Domain-specific terms. If the draft uses "tarball," keep it. Don't sanitize jargon.
  • The author's voice. If the original has a quirk that's intentional (a recurring metaphor, a specific tone), preserve it.
  • Markdown structure when load-bearing — code blocks, headings, links.

Modes

--strip (default)

Edit the draft in place. Apply every category of fix. Save and report the edits.

--detect

Don't edit. Scan the draft and report flagged passages with category + suggested replacement. Let the user accept/reject per item. Useful for high-stakes prose where you want human approval on each change (PR descriptions, blog posts going public, official docs).

Project-aware mode (always on)

Before applying any edits, read the project's CLAUDE.md / AGENTS.md / contributor docs for project-specific style guidance. Honor:

  • Project-specific term preferences ("tarball" not "archive"; "skill" not "plugin")
  • Voice / tone constraints stated in the doc
  • Forbidden phrases the maintainer has called out
  • Any explicit "preserve emdash usage in X context" overrides

If the project has a .deslop.yaml (or similar) configuration file at the repo root, that takes precedence over generic rules.

Workflow

  1. Read the draft.
  2. List the slop categories you spot (emdash, formulaic contrast, opener, connector, bullet explosion, trailing summary).
  3. Apply edits per category in one pass.
  4. Re-read. The text should sound like a person who knows the subject and respects the reader.
  5. If you stripped something the user might want back, mark it: "I removed X — restore if intentional."

Hard rules

  • Don't add new content while deslopping. No new examples, no new caveats. Pure subtraction (or word-level substitution).
  • Don't reduce technical precision. "We pin to a SHA" is precise. "We anchor to a fixed point in history" is slop. Preserve the former.
  • Don't strip emdashes that are doing real work. Where an emdash carries the weight of a thought-pivot, leave it.
  • Don't deslop other people's published writing. This skill operates on drafts. Edited published material has its own choices.

Provenance

Pattern catalog from Jeffrey Emanuel's "Agentic Coding Flywheel" de-slopification guidance. kaijutsu adopts it as a primitive so the catalog lives in one place.

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.