A Khazix-style guide for Chinese WeChat long-form writing about AI, combining supplied material and human observations with expansion, structure suggestions and a four-layer quality report.
Tasks
Understand supplied materials and assess the topic with Happy, Knowledge and Resonance criteria; a passing topic has at least two and an S-grade topic has all three. Ask for missing material or adjust a weak direction.
Choose among investigation/experiment, product experience, phenomenon analysis, tool sharing and methodology patterns, then expand the human-selected angle with evidence, analogies and structure.
Draft a typically 4,000–8,000 Chinese-character WeChat article with short paragraphs, concrete tool names, natural transitions, cultural references and callbacks; mark illustration positions without fabricating experience.
Run L1 hard rules, L2 style consistency, L3 content quality and L4 human-feel review, repair failures and output the prescribed structured quality report before human finalization.
Inputs
Product briefs, news links, PDFs, voice transcripts, scattered ideas, an opening or outline.
The human author’s selected angle, real experience, feelings and verified details.
The referenced style_examples.md and content_methodology.md when more detailed style and topic guidance is needed.
Outputs
A long-form Chinese WeChat draft or expansion in the documented personal style, with illustration markers where needed.
A concise Markdown quality report using the source template: L1 banned words/punctuation/structural clichés/vague tool names; L2 opening/rhythm/colloquial count/punctuation recheck; L3 support/knowledge delivery/cultural connection/empathy/type-specific checks/sequential examples; L4 warmth/uniqueness/posture/flow.
Per-layer pass/fail results, counts or concrete problem locations, the overall result and the top one to three repair priorities.
Limitations and checks
Not for short social posts or title-only summaries. The source’s author context and references focus on AI-industry content; arbitrary-topic expertise is not established.
Do not invent first-person experiences or hypothetical anecdotes presented as real. Admit an untested experience rather than manufacturing it.
The source prohibits its listed Chinese phrases, colons, em dashes and double quotes in the article, and limits structured clichés and excessive bullets/bold blocks. These writing rules do not change the separate Markdown report format.
Only all four quality layers passing makes the article ready under this guide; the framework cannot mechanically guarantee originality or the reader’s subjective response.
Does the topic pass at least two HKR criteria, with a concrete angle and sufficient real material?
Does L1 have zero banned-expression, punctuation, structural-cliché or vague-tool-name hits?
Does L2 preserve the concrete opening, varied rhythm and at least three standalone short paragraphs, with at least 8–10 different colloquial expressions and the punctuation ban rechecked?
Does L3 ground claims in concrete support, integrate knowledge naturally and pass the applicable cultural, empathy, pattern and sequential-example checks?
AI Search
Find projects, verify facts, compare options, or turn a complex need into an actionable plan
Try a searchA click only fills the search box; you stay in control