khazix-skills/khazix-writer
数字生命卡兹克（Khazix）的公众号长文写作skill。当用户需要撰写公众号文章、写稿子、续写文章、根据素材产出长文时使用。触发词包括但不限于：写文章、写稿子、帮我写、续写、扩写、公众号文章、长文、出稿、按我的风格写。即使用户只是说"帮我把这个写成文章"或"用我的风格写一下"，只要上下文涉及内容创作和公众号输出，都应该触发。也适用于用户丢过来一个PDF、brief、新闻链接、语音转文字或任何素材说"帮我写篇文章"的场景。不要用于短内容（小红书帖子、推特、朋友圈）或纯标题摘要生成（那个用wechat-title skill）。
85
85%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Advisory
Suggest reviewing before use
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an excellent skill description that clearly defines its scope (WeChat public account long-form writing), provides extensive natural trigger terms in Chinese, explicitly describes when to use and when NOT to use it, and even references a related skill by name to prevent conflicts. The only minor note is that it's written in a somewhat verbose style, but the verbosity serves the purpose of disambiguation and trigger coverage.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple concrete actions: writing WeChat public account long-form articles, continuing/expanding articles, producing long-form content from materials (PDF, briefs, news links, voice-to-text). Also specifies what it does NOT do (short content, title/summary generation). | 3 / 3 |
Completeness | Clearly answers both 'what' (WeChat public account long-form article writing, continuation, expansion from materials) and 'when' (explicit trigger words listed, usage scenarios described, and negative boundaries defined with what NOT to use it for). The explicit 'Use when' equivalent and exclusion clauses are thorough. | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural trigger terms in Chinese: 写文章, 写稿子, 帮我写, 续写, 扩写, 公众号文章, 长文, 出稿, 按我的风格写. Also includes natural user phrases like '帮我把这个写成文章' and '用我的风格写一下', covering many variations users would actually say. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive: explicitly scoped to 公众号长文 (WeChat long-form articles) for a specific persona (数字生命卡兹克), and explicitly excludes short content platforms (小红书, Twitter, 朋友圈) and title/summary generation (referencing wechat-title skill by name). This makes conflict with other skills very unlikely. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
62%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a deeply detailed and highly actionable writing style guide with excellent workflow structure and a rigorous four-layer self-check system. Its main weakness is extreme verbosity — it spends thousands of tokens explaining writing theory, narrative psychology, and creative philosophy that Claude already understands, when it could instead focus on the concrete rules, examples, and checklists that make it genuinely useful. Splitting the content across reference files and trimming explanatory prose would significantly improve token efficiency.
Suggestions
Move the lengthy explanations of writing techniques (Hero's Journey, Chekhov's Gun, 升番 comedy theory, Gestalt psychology references) into a separate references file, keeping only the actionable rules and examples in the main SKILL.md
Condense the style descriptions by removing 'why it works' explanations and keeping only 'what to do' instructions — e.g., instead of explaining why 回环呼应 works with a paragraph about Chekhov's Gun theory, just say 'Plant hooks early, callback to them later' with one example
Extract the L1-L4 self-check system into a separate checklist file (e.g., references/self_check.md) and reference it from the main skill, reducing the main file's length by ~40%
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~4000+ words. It extensively explains writing philosophy, narrative theory (Hero's Journey, Chekhov's Gun), comedy structure (升番), and creative principles that Claude already understands. Much of this could be condensed to concrete rules and examples rather than lengthy explanations of why each technique works. | 1 / 3 |
Actionability | The skill provides highly specific, actionable guidance: exact banned words with replacements, specific punctuation rules, concrete phrase lists to use, detailed checklist items with pass/fail criteria, structural templates, and clear examples of good vs bad patterns. A writer could follow these instructions precisely. | 3 / 3 |
Workflow Clarity | The four-step workflow (understand material → clarify AI role → write → self-check) is clearly sequenced. The L1-L4 self-check system has explicit pass/fail criteria, specific repair instructions, and a structured output format. The human-AI collaboration flow is also clearly diagrammed with validation loops. | 3 / 3 |
Progressive Disclosure | The skill references two external files (references/style_examples.md and references/content_methodology.md) at the end, which is good. However, the main file itself is monolithic with enormous amounts of inline content that could be split into separate reference files (e.g., the full banned words list, the oral phrase recommendations, the detailed L1-L4 checklist could each be separate files). The structure within the file is reasonable but the sheer volume hurts navigability. | 2 / 3 |
Total | 9 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
Reviewed
Table of Contents