← 返回 Skills 市场
wechat-style-writer
作者
harven-droid
· GitHub ↗
· v1.0.3
· MIT-0
391
总下载
2
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install wechat-style-writer
功能描述
模仿指定老师/作者的写作风格改写公众号文章。当用户说"用XX风格改写"、"模仿XX写一篇"、"按XX老师的风格重写"、"帮我改成XX的风格"、或上传文章要求风格改写时触发。也适用于用户提到"风格模仿"、"文章改写"、"公众号改写"等场景。即使用户只说"帮我改写这篇文章"也应触发此技能,因为改写是它的核心功能。
安全使用建议
This skill appears to do exactly what it says: rewrite user-provided articles into the writing styles stored under references/styles/. Before installing or using it, consider: (1) legal/ethical risk—mimicking living authors can raise impersonation or copyright issues even though SKILL.md instructs removal of personal identifiers; avoid publishing outputs that could be mistaken for the original author or that reproduce copyrighted full-texts from the style samples; (2) privacy/local storage—the skill can save .md files to /mnt/user-data/outputs when you confirm, so check that location and delete unwanted files; (3) provenance of style samples—the repository includes full sample articles for each author (these may be copyrighted), so ensure you have rights to store/use them; (4) safety—there are no network calls, no secret requirements, and no installs declared. If you proceed, review generated content before posting publicly and consider adding explicit user consent and a copyright/attribution policy for using author styles.
功能分析
Type: OpenClaw Skill
Name: wechat-style-writer
Version: 1.0.3
The skill bundle is a text-processing utility designed to rewrite articles in the style of specific Chinese bloggers. It operates by reading local style profiles from 'references/styles/' and writing Markdown outputs to '/mnt/user-data/outputs/'. While the instructions in 'SKILL.md' include techniques to evade AI detection (such as simulating human typos and removing original author identifiers), these are functional requirements for the stated purpose of 'style imitation' and do not constitute a security threat, data exfiltration, or unauthorized system access.
能力评估
Purpose & Capability
Name/description (wechat-style-writer) match the actual behavior: it reads local style profiles and a user-supplied article and rewrites the article in that style. No unrelated env vars, binaries, or install steps are requested.
Instruction Scope
SKILL.md instructs the agent to read style files under references/styles/ and any user-uploaded text/docx/xlsx, perform generation and an internal 'AI-flavor' self-check, then display results. These steps stay within the stated purpose. Two user-impacting items to note: (1) it explicitly directs removal/replacement of authors' personal identifiers (good) but nevertheless enables close imitation of living authors (legal/ethical risk outside of scope of pure coherence); (2) it instructs saving output to /mnt/user-data/outputs and to use a present_files tool for download, which affects local storage and should be disclosed to users.
Install Mechanism
Instruction-only skill with no install spec and no code files to execute. This minimizes disk/execution risk; nothing is downloaded or installed.
Credentials
The skill requires no environment variables, credentials, or external endpoints. The declared resources (local style files and user uploads) are proportionate to the stated task.
Persistence & Privilege
always:false and normal autonomous invocation default. The only persistent effect is optionally writing .md outputs to /mnt/user-data/outputs and using present_files to expose them; this is expected for a content-generation skill but the user should be aware files will be created on the agent's filesystem if they confirm saving.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install wechat-style-writer - 安装完成后,直接呼叫该 Skill 的名称或使用
/wechat-style-writer触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.3
wechat-style-writer 1.0.3
- 增加了“正文分段加小标题”流程,改写时自动按目标老师习惯为每个段落加合适的小标题或编号
- 详细说明了各主流老师(如卡兹克、半佛、大瑜、老徐)段落分隔和小标题的具体风格规则
- 新增“AI味自检”环节,明确输出前需与风格范文对比,自动检查语气、节奏、分段、表达和AI痕迹
- 明确要求根据“段落与结构”风格调整分段数量和小标题,不可生搬硬套,须自然拆分
- 其他执行流程未改动,功能核心保持一致
v1.0.2
wechat-style-writer 1.0.2
- 增加“去个人标识”规范:改写时移除所有原作者的个人信息、署名、联系方式、专属社群等内容,仅保留风格骨架,保障模仿仅限写作风格而非身份冒用。
- 结尾部分调整:结尾的个人信息全部用通用占位符(如“[你的公众号名称]”等)替换,且禁止使用原作者自称。
- 输出流程优化:改写结果先展示整页 Markdown,用户确认后再选择是否保存、下载,提升交互体验。
- 补充规范:强化仿写风格不等同冒充原作者、不得出现运营抽奖等原作者专属内容。
- 其它描述补充,提升易用性和安全性。
v1.0.1
删除中文
v1.0.0
wechat-style-writer 1.0.0 – 首次发布!
- 支持用户上传或粘贴公众号原文,指定目标老师/作者,自动改写为其写作风格。
- 根据指定风格档案,模拟标题、结构、语气、错别字与排版习惯,生成高仿真“仿写”Markdown文档。
- 一键输出推荐+备选3个风格标题,正文完整、支持常见错别字/笔误以减少AI感。
- 支持自动列出可用风格老师,并指引如何添加新风格档案。
- 生成结果自动保存在 outputs 文件夹,并展示给用户。
元数据
常见问题
wechat-style-writer 是什么?
模仿指定老师/作者的写作风格改写公众号文章。当用户说"用XX风格改写"、"模仿XX写一篇"、"按XX老师的风格重写"、"帮我改成XX的风格"、或上传文章要求风格改写时触发。也适用于用户提到"风格模仿"、"文章改写"、"公众号改写"等场景。即使用户只说"帮我改写这篇文章"也应触发此技能,因为改写是它的核心功能。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 391 次。
如何安装 wechat-style-writer?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install wechat-style-writer」即可一键安装,无需额外配置。
wechat-style-writer 是免费的吗?
是的,wechat-style-writer 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
wechat-style-writer 支持哪些平台?
wechat-style-writer 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 wechat-style-writer?
由 harven-droid(@harven-droid)开发并维护,当前版本 v1.0.3。
推荐 Skills