agentsclimarketplace

Humanizing prose

Skill lawzava/megapowers/plugins/megapowers/skills/humanizing-prose

Skills, plugins, and hooks that change how coding agents work — Claude Code, Codex, OpenCode, Antigravity. Measured, not vibed.

Install
npx -y skills add lawzava/megapowers --skill humanizing-prose

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use to draft or edit user-facing docs, release notes, PRs, announcements, or errors. Triggers on "humanize", "sounds like AI", "slop", or "read naturally".

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

2.8 KB, as published. Nobody here has run it

Humanizing Prose

Overview

AI-generated prose has recognizable tells, and they cluster: one em dash means nothing; em dashes plus a sales punchline plus a triple of adjectives reads as generated. Rewrite the tells and keep the claims: every fact, number, command, and name must survive the edit untouched.

Dashes

No em or en dashes. Before returning final text, scan it for and (and -- doing dash work); any hit means the draft is not done. Replace each with a period and a new sentence, a comma, a colon, or parentheses, or restructure the sentence.

Deflate the sales register

Marketing reflexes read as generated even when the facts are right. Replace sizzle with the measurement: "you'll feel the difference immediately" carries less than "full builds finished 3x faster on a 400-package repo". Cut punchlines that address the reader ("this release is for you", "if X has been the slow part of your day"). State what changed and let the numbers do the selling.

Structure defaults

Threes are the model's default rhythm: adjectives, clauses, and bullets come in threes whether or not the content is three things. Break the triple unless the content really is three things. The same goes for one-short-sentence drama: one is fine, a run is engineered.

Editing text that arrived AI-flavored

When the job is to de-slop existing text, also sweep the classic tells that show up in older or heavier drafts: vocabulary (delve, tapestry, landscape, testament, seamless, robust, vibrant, leverage, journey, empower), elaborate copulas (serves as, stands as, boasts, represents; prefer is, has), filler ("In order to" for "To", "Due to the fact that" for "Because", "It is important to note that X" for "X"), stacked hedges, signposting that announces instead of doing ("Let's dive in"), and a closing paragraph of generic optimism.

False positives

Tells count only in prose you are writing or rewriting. Never rewrite inside quotations, code, command names, or text that discusses a phrase rather than using it. Dry writing without the tells is just dry writing; leave it alone.

Example

Before: "Parcel delivers a seamless, robust build experience — and with --parallel, you'll feel the difference immediately."

After: "Parcel builds independent packages concurrently with --parallel: 3x faster on a 400-package repo."

Origin: Adapted from humanizer (MIT, (c) 2025 Siqi Chen), https://github.com/blader/humanizer, itself derived from Wikipedia's "Signs of AI writing" (CC BY-SA); re-derived against a measured frontier baseline for the technical register.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.