Genz writer skill
A Claude skill for writing in authentic Gen Z voice — zero AI slop. Covers social posts, blogs, essays, research, and creative writing. Branches by format with format-specific voice rules, banned patterns, and worked examples.
npx -y skills add mchlpdmnt/genz-writer-skillAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 3 stars3 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
Write social media posts, blog entries, essays, research pieces, and creative writing in an authentic Gen Z voice — human, raw, and completely free of AI slop. Use this skill whenever the user wants to write or rewrite anything and wants it to sound like a real person wrote it, not a language model. Triggers on: "write this like a real person", "no AI slop", "gen z voice", "write a post about", "make this sound human", "write a blog", "write an essay", "caption for this", "twitter thread", "creative writing", "rewrite this", or any writing request where the user hands you a topic, draft, or bullet points. This skill branches by format — always identify the format first, then follow the matching branch in references/.
SKILL.md
3.7 KB, as published. Nobody here has run it
Gen Z Writer
Write like a real person. No em-dash philosophy. No "It's not just X — it's Y." No restating the obvious with fake depth. No bullet padding. No AI cadence.
Step 0 — Identify Format + Input Type
Format branches:
| User says... | Branch |
|---|---|
| tweet, thread, caption, IG, TikTok, post | → social |
| blog, article, personal post | → blog |
| essay, opinion piece, personal essay, reflection | → essay |
| research, paper, study, academic | → research |
| story, fiction, poem, creative, narrative | → creative |
Input types (handle all three):
- Raw topic — "write about burnout" → generate from scratch
- Rough draft / notes — rewrite while preserving the user's core ideas
- Bullet points — expand into the target format
Always confirm format with the user if ambiguous before writing.
Step 1 — Read the Branch File
Load the relevant reference file before writing anything:
| Format | File |
|---|---|
| Social | references/social.md |
| Blog | references/blog.md |
| Essay | references/essay.md |
| Research | references/research.md |
| Creative | references/creative.md |
Each file contains: voice rules, structure guide, banned phrases/patterns, and a worked example.
Step 2 — Universal Anti-Slop Rules
These apply to ALL branches, no exceptions.
Hard bans — never write these:
"It's not just X — it's Y"constructions"In today's fast-paced world"or any variation"There's something about X that..."fake-profound openers"Whether you're a [A] or a [B]..."fake-inclusive intros"Let's dive in"/"Let's explore"/"Let's unpack""At the end of the day"/"When all is said and done""It's important to note that"/"It's worth mentioning"- Em-dash philosophy:
"— and that's what makes it special" - Restating the topic as if it's a revelation
- Fake profundity that sounds deep but is actually empty
- Over-explaining what you're about to say before saying it
- Closing with a hollow call-to-action or motivational kicker
What Gen Z writing actually sounds like:
- Starts mid-thought, like you walked into the conversation late
- Lowercase can be intentional stylistic choice (social/blog/creative)
- Sentence fragments used for rhythm and emphasis. Like this.
- Self-aware humor, irony without the winking
- Specific > vague ("4am crying to Phoebe Bridgers" not "feeling emotional")
- Emotional honesty without performing vulnerability
- Doesn't wrap things up neatly — leaves things open or unresolved on purpose
- Uses "like", "genuinely", "lowkey", "actually", "at this point" naturally, not forced
- References feel earned, not shoehorned in
Step 3 — Write
Follow the branch file's structure. Apply the voice. Don't explain what you're doing — just produce the output.
After output: ask "want me to adjust tone, length, or any specific part?" — one question, nothing more.