What does this skill do, and when should you use it?
This Skill enables AI agents to write WeChat public-account long-form articles in the distinctive style of 'Digital Life Khazix' (a Chinese AI industry observer). It encodes a comprehensive writing methodology: five article archetypes, colloquial phrasing library, opening techniques, escalation logic, Chekhov's gun, and more. A four-layer quality check (hard rules, style consistency, content depth, human-feel final review) helps avoid AI-sounding output. It emphasizes human-AI collaboration: AI provides evidence, analogies, and expansion, while humans supply core insights and real experiences. The Skill explicitly rejects clichés, banned words, and textbook-style openings, making it ideal for creators who want authentic personal narratives.
The Skill takes user-provided source materials (PDF, news links, voice transcripts, scattered ideas) and first evaluates the topic using the HKR framework (Happy, Knowledge, Resonance). It then structures content according to one of five article archetypes. It applies Khazix's writing style rules: conversational pacing, deliberate discourse breaks, knowledge output style, personal perspective, cultural elevation, and callback loops. After drafting, it runs a four-layer self-check report (L1 banned word/punctuation scan, L2 style consistency, L3 content quality, L4 human-feel final review) and outputs revision suggestions. It also provides reference files (style_examples.md and content_methodology.md) for the agent to consult.
- A WeChat public-account owner wants an industry analysis article in Khazix's style, handing a PDF brief to the agent
- A content creator needs to turn a voice transcript into a stylistically consistent WeChat article
- An editor wants to continue or expand an existing draft while maintaining voice consistency
- A startup founder writing a product experience article wants the 'come play with me' narrative structure
- A methodology enthusiast wants to share systematic insights with 'humble opening' and actionable steps
How do you install this skill?
- Depends on external reference files (style_examples.md, content_methodology.md); ensure they are distributed with skill, else instructions incomplete.
- No tests or execution evidence; actual output quality unverified.
- User material may contain sensitive info; skill lacks privacy protections.
- If relying on WeChat platform, network reachability may be an issue.
- Local filesystem
In an Agent Skills-compatible tool (e.g., Claude Code, Codex), say: 'Install this skill: https://github.com/KKKKhazix/khazix-skills/tree/main/khazix-writer'. The agent will clone it to the appropriate directory. If your tool doesn't support the Skills standard, download the SKILL.md file and use it as a project rules file.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/KKKKhazix/khazix-skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/khazix-writer" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Write a WeChat long-form article
After installation, instruct the agent to 'Write a WeChat long-form article' or 'Write this material in Khazix's style', providing source material (PDF, links, bullet points, etc.). The agent will first confirm the topic direction and may ask for more details. Follow the collaborative workflow: you provide core ideas and experiences, AI provides evidence and expansion. After drafting, review the four-layer quality report before finalizing.
What are this skill's strengths and limitations?
- Extremely detailed style rules covering vocabulary, punctuation, and narrative structure, effectively reducing AI-sounding output
- Built-in four-layer self-check system with clear pass/fail criteria and fix suggestions
- Emphasizes human-AI collaboration to avoid fully generated content that lacks authenticity
- Explicitly rejects clichés and banned words, suitable for writers seeking genuine voice
- Accepts various input formats (PDF, links, transcripts, etc.), flexible to use
- Tailored for Chinese WeChat long-form content, not suitable for short-form or English writing
- Style is highly personal; not applicable if you don't like Khazix's style
- Requires active user participation for core insights and experiences; cannot run fully automatically
- No automated tests or validation scripts; relies on human final review
- May have a learning curve for users new to Agent Skills
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Khazix WeChat Long-Form Writer Skill this page | 49 · Use with care | ★ 21k | 10d ago | MIT |
| e2e Docs Authoring Skill | 54 · Use with care | ★ 8.7k | 1d ago | Apache-2.0 |
| No AI Slop | 59 · Recommended | ★ 12k | 1mo ago | MIT |
| Beautiful Prose Style Contract | 50 · Use with care | ★ 57 | 9mo ago | — |
| Unslop | 49 · Use with care | ★ 8.7k | 1d ago | Apache-2.0 |
This Skill is a purpose-built personal style generator for WeChat long-form articles, distinct from general-purpose writing assistants. It explicitly rejects standardized expressions and requires genuine human-AI collaboration. Compared to other AI writing tools, it prioritizes 'human feel' and style consistency over generic eloquence.
How did FollowSkills review this skill?
This skill is a pure text instruction set, no privilege escalation, external service calls, or sensitive data collection. It asks users for raw material and outputs articles, data flow is transparent. No hidden network requests or file operations, low dependency risk. But it does not clearly state how user material privacy is handled, nor does it have rollback or recovery mechanisms, so points deducted.
Instructions are detailed and self-consistent, but rely on external files (style_examples.md, content_methodology.md) not provided in SKILL.md, and no tests, error handling, or failure feedback are included. Static review cannot verify actual run, so max 7.
Clear description, specific triggers, suited for Chinese WeChat long-form. But no explicit non-fit scenarios, e.g., non-Chinese or short content, and environment fit depends on WeChat platform, which may be reachable in mainland China, but external links (mp.weixin.qq.com) may be restricted. Triggers good but boundaries slightly vague, so deducted.
SKILL.md well-structured, has examples and references, but lacks version, changelog, maintenance responsibility, and install notes. License exists but not stated in SKILL.md. Overall readable and usable, but versioning and governance incomplete, so 9.
Cannot execute to verify output quality. Instructions include self-check, but actual effect needs dynamic verification. Static review cannot confirm task completion, so 5.
No tests, CI, or third-party verification. Claims reference files but no tests provided, style examples are self-authored, lack independent verification, so 2.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →