isatimur/de-slop / skills/de-slop
de-slop
VerifiedUse when prose reads like AI — to remove "AI slop" (empty hedging, listicle stems, smooth transitions that hide the absence of a claim, generic filler) and rewrite it into writing with a real point of view. Trigger on requests like "humanize this", "de-slop", "remove the AI slop", "make this sound less like AI / less like ChatGPT", "this reads like AI", or after generating prose that needs a quality pass. Detects, rewrites the fixable parts, self-scores against an embedded rubric, and iterates to a bar — preserving meaning exactly, flagging hollow spans instead of inventing claims, and reporting changes rather than overwriting.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.4 KB |
| references/examples.md | markdown | 8.6 KB |
| references/guardrails.md | markdown | 2.4 KB |
| references/rubric.md | markdown | 3.8 KB |
| references/slop-catalogue.md | markdown | 9.9 KB |
| scripts/build_adapters.py | script | 7.6 KB |
| scripts/flag_slop.py | script | 21.2 KB |
| scripts/setup-github.shexecutable | script | 2.0 KB |
references/rubric.md
View on GitHub# Humanness Rubric
Score a single paragraph on **humanness** — does this sound like a thinking
author with a point of view, or is it generic AI-flavored prose that could appear
in any blog post?
## Bands (0–100)
- **strong (80–100):** specific, has a point of view, survives a hostile editor's
red pen. A real claim someone could disagree with. Removing it would lose
something.
- **moderate (50–79):** readable but generic in spots. A real point is present
but softened by hedging or filler.
- **weak (20–49):** pattern-matchable AI prose. Removing it loses nothing. The
sentence is shaped like an argument but isn't making one.
- **fail (0–19):** pure scaffolding language — listicle stems, empty hedging,
transitions with no content between them.
## Slop indicators (lower the score)
- **Empty hedging:** "it's worth noting", "it's important to remember", "that
said", "of course", "arguably" used to avoid committing.
- **Listicle stems with no point of view:** "There are several key factors…",
"Here are a few things to consider…" followed by the obvious.
- **Smooth transitions that hide the absence of a claim:** "Moreover", "In
addition", "Furthermore" gluing together sentences that don't actually advance
an argument.
- **Generic filler:** intensifiers ("really", "very", "truly", "incredibly"),
manufactured stakes ("in today's fast-paced world"), throat-clearing openers.
For the full taxonomy of tells — each one, why it reads as AI, and which detector
type catches it (or why none can) — see `slop-catalogue.md`. Note that the
detector surfaces *candidates* by surface pattern; this rubric assigns the band.
A clean-looking paragraph can still be **weak** or **fail** if it makes no claim
(hollowness is invisible to any regex — only the removal test below catches it).
## The two tests
1. **Hostile-editor test:** would a sharp editor leave this sentence on the page,
or red-pen it as padding?
2. **Removal test:** if you deleted this sentence, would the reader lose
anything? If nothing is lost, the sentence is slop regardless of how polished
it sounds.
## The substance lens (specificity · restraint · voice)
Surface tells are only half the judgment. A paragraph can be mechanically clean —
zero hedging, no buzzwords — and still be slop because it has nothing to say. When
the detector is quiet, judge substance on three axes before passing it:
- **Specificity** — does it commit to a concrete claim (a name, number, mechanism,
consequence), or only gesture at one? "Significant improvement" fails;
"latency dropped from 340ms to 90ms" passes.
- **Restraint** — is the emphasis earned, or manufactured? Forced contrast, dramatic
fragmentation, and hot takes are *negative* substance — louder, not stronger.
- **Voice** — is there a thinking author reacting to the facts, or a neutral
narrator restating them? Voice is not personality theatrics; it is having a point.
Mechanics-clean ≠ good. This lens is what separates *moderate* from *strong*, and
it is the human/model side of the loop the regex cannot reach. (The three-axis
"substance" framing is shared with `stop-slop` derivatives like `tagore`; here it
guides band judgment rather than a numeric gate.)
## Triage rule (for anything below strong)
- **REWORDABLE** — the paragraph *has a real claim* that's buried under hedging or
filler. The fix is subtraction + sharpening. → rewrite it.
- **HOLLOW** — the paragraph is weak because it *has no actual point to make*.
Nothing is lost if removed. Rewording cannot fix an absent claim. → flag it;
do not invent a claim to fill the hole.
The single most important judgment call this skill makes is rewordable vs.
hollow. When unsure, apply the removal test: if deleting the paragraph entirely
would cost the reader nothing, it is hollow.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |