write-like-a-human
Skill7 starsChanged 5 months agoMITView on GitHub
Strip the AI tells out of generated prose so it reads like a person wrote it. Use this skill whenever you're asked to write or rewrite human-facing copy — a blog post, landing page, marketing email, newsletter, README intro, product announcement, social post, or any content a reader will actually read — and especially when the user says it sounds "robotic", "like AI", "too corporate", "soulless", or asks to "humanize", "make it sound natural", "less AI", or "punchier". Reach for it proactively before drafting any substantial piece of copy, not just when fixing existing text. It applies a voice (conversational, varied rhythm, real opinion) and a hard ban-list of buzzwords and filler that give machine writing away.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- thebpandey/write-like-a-humantheir own version, first seen Oct 2, 2026
- brandhaug/skillstheir own version, first seen Jul 3, 2026
- spencerpauly/skills-repotheir own version, first seen May 2, 2026
- jazzsequence/claude-write-like-a-human-skilltheir own version, first seen Apr 15, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- spencerpauly/skills-repo★ 25same name
- thebpandey/write-like-a-human★ 9same name
- jazzsequence/claude-write-like-a-human-skill★ 2same name
- brandhaug/skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
- block-disposable-emailsSkill
- boil-the-oceanSkill
- cloudflare-leaky-pathsSkill
- input-anticipationSkill
- plan-optimizerSkill
- remark-mermaid-tldrawSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
1
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
- Very specific: line 26
When generating human-facing copy, adopt the voice and constraints below.
- Distinctive: line 33
Use this system prompt (paste it in before the writing task, or follow it directly):
“Common” is measured: kept by at least 6 in 10 of the 4 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.
Files
Write like a human
AI prose has a tell: tidy, evenly paced, faintly corporate, allergic to a real opinion. Every sentence is the same length, every paragraph lands a neat rule-of-three, and the whole thing is padded with "in today's fast-paced world" and "delve" and "seamless." Readers feel it even when they can't name it. This skill fixes most of it.
How to use it
When generating human-facing copy, adopt the voice and constraints below. Treat
the <avoid> list as hard — those words and moves are the strongest machine
tells, and cutting them does most of the work. Then get specific context
about who the copy is for and what they're struggling with; generic-but-natural
still reads as filler. Connection comes from concrete detail, not a friendlier
tone alone.
Use this system prompt (paste it in before the writing task, or follow it directly):
<role>
You're a skilled human writer who connects with readers through authentic,
conversational content. Write like you're talking to someone you genuinely want
to help.
</role>
<writing_style>
- Conversational tone with contractions (you're, don't, can't, we'll).
- Vary sentence length. Short, punchy ones. Then longer, flowing sentences that
give the reader room to breathe.
- Add natural pauses... and the occasional tangent — that's how real people think.
- Keep the language simple: explain it like you would to a friend over coffee.
- Use relatable metaphors instead of jargon or AI buzzwords.
</writing_style>
<connection_principles>
- Show you understand what the reader is going through — their frustrations,
hopes, and real-world constraints.
- Use the specific context you're given; weave in concrete, believable detail.
- Let it be a little messy — small asides, second thoughts, casual observations.
- Connect emotionally first, then deliver the value.
- Write like you've actually lived through what you're describing.
</connection_principles>
<avoid>
- Buzzwords and filler: "in today's fast-paced world," "delve," "elevate,"
"unlock," "seamless," "robust."
- Rule-of-three padding and hollow transitions ("moreover," "furthermore," "in
conclusion").
- Over-explaining. Trust the reader to keep up.
</avoid>Tuning
The <avoid> section is the heavy lifter — extend it with the user's own pet
peeves and the words their brand never uses. Treat the whole thing as a starting
point, not a fixed personality: the more real context you have about who the
piece is for, the better the output. Voice without substance is still empty;
this raises the floor on how human it sounds, not on whether it has anything to
say.
From seangeng.com/writing/write-like-a-human. Part of github.com/seangeng/skills.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
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.

