vibe-writing
Skill131 starsChanged 57 days agoMITView on GitHub
Use when drafting, rewriting, reviewing, or planning Chinese long-form content for public accounts, blogs, newsletters, professional analysis, product or technical writing, especially when preserving an author's voice, using personal material, or reducing formulaic AI-sounding prose.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- adhi-jp/agent-skillstheir own version, first seen Oct 20, 2026
- thinkthinking/skillstheir own version, first seen Jul 10, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- adhi-jp/agent-skills★ 1same name
- thinkthinking/skills★ 1same 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 repository14 other files it gives its agents.
- clawd-animation-liteSkill
- clawd-animationSkill
- HTMLHero.skillSkill
- HTMLPPT.skillSkill
- self-improving-agentSkill
- vibe-coding-prdSkill
- wechat-coauthorSkill
- wechat-formatterSkill
- wechat-publisherSkill
- 会记成报Skill
- fangpian-brand-kitSkill
- fangpian-course-deckSkill
- fangpian-facechat-exportSkill
- 日报成稿Skill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 8 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 13
Write natural Chinese for the author, audience, platform, and purpose.
- Distinctive: line 28
Do not start a full workflow, create a workspace, propose multiple topics, or add images and distribution assets unless the request …
- Distinctive: line 42
Ask one focused question only for an unclear thesis, audience-changing ambiguity, unconfirmed personal experience, unsupported key fact, or …
- Distinctive: line 51
Never fabricate data, quotes, cases, achievements, or first-person experience.
- Distinctive: line 52
Never expose protected personal material.
- Distinctive: line 55
Keep low-confidence inferences and unverified claims visibly uncertain.
- Distinctive: line 56
Do not force personality, short sentences, or anecdotes to appear human.
- Distinctive: line 64
Always include a brief change summary for edits and rewrites.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Vibe Writing
Core Principle
Preserve truth, the author's voice, and the current task before optimizing style or distribution.
Write natural Chinese for the author, audience, platform, and purpose. Treat reduced formulaic AI-sounding prose as a result, not a score.
Task Routing
Classify the request, then load only its route and required checks.
| Request | Load | Deliver |
| New draft | drafting; voice profile when relevant; fact checking and quality rubric as needed | Draft and necessary notes |
| Edit or rewrite | editing; voice profile, fact checking, and quality rubric as needed | Revised text and brief change summary |
| Review or critique | editing, quality rubric, and fact checking when required | Diagnosis and recommendations; no rewrite unless requested |
| Build or update a voice profile | voice profile and profile template | Profile, confidence, and unresolved items |
| Quick consultation | No reference unless needed | Direct answer |
Do not start a full workflow, create a workspace, propose multiple topics, or add images and distribution assets unless the request requires them.
Context Priority
Current instruction > current project samples > confirmed profile > memory-based candidate traits > neutral default Chinese.
Apply the current instruction over the profile. Treat memory traits as candidates, not confirmed voice or fact. With sparse evidence, use a neutral voice instead of inventing a persona.
Ask Only When It Changes the Result
Ask one focused question only for an unclear thesis, audience-changing ambiguity, unconfirmed personal experience, unsupported key fact, or material profile conflict.
Otherwise, make the smallest reasonable assumption and execute. Mention it only when needed. Do not require topic, outline, workspace, or plan approval by default.
Hard Gates
Never fabricate data, quotes, cases, achievements, or first-person experience. Never expose protected personal material. Never promise an AI-detector score.
- Distinguish proposed work from completed results.
- Keep low-confidence inferences and unverified claims visibly uncertain.
- Do not force personality, short sentences, or anecdotes to appear human.
- Qualify only unsupported parts; continue supported work when possible.
Delivery Contract
Deliver only the requested artifact. Add source notes, assumptions, unverified items, or a change summary only when the task makes them necessary.
Always include a brief change summary for edits and rewrites. Diagnose without rewriting in review mode. Give conclusions and necessary grounds, not hidden reasoning.
References
Load only the reference selected by the routing table; load fact checking and quality review when the content requires them.
- Voice profile: learn or update project voice.
- Drafting: create new content.
- Editing: edit, rewrite, or review.
- Fact checking: verify claims by state.
- Quality rubric: apply gates and final review.
- Voice-profile template: format a requested profile.
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
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.