Skip to content

NocodeMrLi/mr-li-writer-skill

mr-li-writer

Verified

面向中文内容创作的编辑判断、研究、策划、写作与平台原生交付工作流。根据标题、素材、参考资料、核心想法或模糊灵感,先识别读者任务、材料条件和创新必要性,再按需完成语境扫描、可写性判断与方向校准,并区分发布平台与内容目标,进行事实检索、观点设计、平台原生正文写作、标题策划和对应交付。适用于公众号文章、知乎回答、小红书原生笔记、官网/网页、个人博客、行业解读、实用指南和基于素材的扩展研究写作。

FileKindSize
SKILL.mdexecutableskill md62.0 KB
agents/openai.yamlexecutablescript12.9 KB
assets/gzh-design/LICENSEexecutablescript34.2 KB
assets/gzh-design/NOTICE.mdmarkdown568 B
assets/gzh-design/assets/preview-template.htmlexecutablescript3.8 KB
assets/gzh-design/references/common-components.mdexecutablescript14.3 KB
assets/gzh-design/references/eval-cases.mdexecutablescript5.7 KB
assets/gzh-design/references/format-normalize.mdexecutablescript4.4 KB
assets/gzh-design/references/theme-generator.mdexecutablescript25.3 KB
assets/gzh-design/references/theme-graphite-minimal.mdexecutablescript34.1 KB
assets/gzh-design/references/theme-index.mdexecutablescript3.1 KB
assets/gzh-design/references/theme-moyu-green.mdexecutablescript46.1 KB
assets/gzh-design/references/theme-moyu-ticket.mdexecutablescript24.7 KB
assets/gzh-design/references/theme-olive-journal.mdexecutablescript56.2 KB
assets/gzh-design/references/theme-red-white.mdexecutablescript29.9 KB
assets/gzh-design/references/theme-zen-whitespace.mdexecutablescript23.5 KB
assets/mr-li-writer-promo-poster.pngbinary179.4 KB
assets/mr-li-writer-promo-silent.mp4binary1.2 MB
assets/readme-banner.svgother5.3 KB
assets/readme-showcase/moyu-green-content.jpgbinary281.1 KB
assets/readme-showcase/moyu-green-cover.jpgbinary183.8 KB
assets/readme-showcase/moyu-green-list.jpgbinary203.4 KB
assets/readme-showcase/moyu-green-long.jpgbinary431.9 KB
assets/readme-showcase/olive-journal-content.jpgbinary177.7 KB
assets/readme-showcase/olive-journal-cover.jpgbinary214.2 KB
assets/readme-showcase/olive-journal-ending.jpgbinary154.1 KB
assets/readme-showcase/olive-journal-long.jpgbinary352.3 KB
assets/readme-showcase/red-white-cover.jpgbinary183.1 KB
assets/readme-showcase/red-white-long.jpgbinary405.3 KB
assets/readme-showcase/red-white-section.jpgbinary227.4 KB
assets/readme-showcase/red-white-table.jpgbinary219.4 KB
references/commercial-source-terms.txtmarkdown389 B
references/delivery-protocol.mdmarkdown11.7 KB
references/editorial-evaluation-protocol.mdmarkdown3.1 KB
references/editorial-routing-protocol.mdmarkdown6.2 KB
references/genre-structure-protocol.mdmarkdown7.5 KB
references/high-impact-writing-protocol.mdmarkdown6.2 KB
references/humanize-rules.mdexecutablescript11.5 KB
references/ideation-protocol.mdmarkdown6.7 KB
references/lint-checks.mdmarkdown3.9 KB
references/opening-design-protocol.mdmarkdown9.2 KB
references/platform-native-protocol.mdmarkdown19.3 KB
references/research-protocol.mdexecutablescript13.9 KB
references/title-rules.mdexecutablescript9.2 KB
references/writing-rules.mdexecutablescript18.4 KB
scripts/build_gzh_html.pyexecutablescript70.3 KB
scripts/build_html.pyexecutablescript31.8 KB
scripts/component_lint.pyexecutablescript8.1 KB
scripts/extract_seed.pyexecutablescript5.8 KB
scripts/gzh_component_inventory.pyexecutablescript1.8 KB
scripts/lint_article.pyexecutablescript19.5 KB
scripts/scan_sensitive.pyscript2.0 KB
scripts/source_governance.pyscript4.9 KB
scripts/validate_delivery_bundle.pyexecutablescript18.3 KB
scripts/validate_gzh_html.pyexecutablescript14.4 KB
scripts/validate_research_scope.pyexecutablescript10.0 KB
scripts/validate_task_intake.pyscript27.3 KB
scripts/wrap_gzh_preview.pyexecutablescript3.7 KB

references/editorial-evaluation-protocol.md

View on GitHub
# 编辑效果评测规范

这套规范用于验证规则升级是否真的让文章更值得读,而不是只让提示词更复杂。它不参与单篇文章的强制生成流程,适合版本升级、规则调整和模型适配后的回归评测。

## 一、评测原则

1. **盲评**:隐藏版本、模型和提示词信息,只展示读者实际看到的标题与正文。
2. **跨体裁**:使用 `tests/fixtures/editorial-routing-cases.md` 中的 12 个种子题,覆盖查询、指南、观点、关系、故事、产品与平台原生内容。
3. **同题多跑**:至少选 4 个种子题各生成 3 次,检查标题、开头、结构和论证路径是否出现机械复用。
4. **基线对照**:保留升级前版本或关闭新路由后的结果,避免只凭主观印象判断进步。
5. **只评读者可见结果**:不因为流程更长、检索更多或术语更专业而加分。

## 二、核心维度

每项 1-5 分,并记录一句具体理由:

| 维度 | 重点问题 |
| :--- | :--- |
| 点击意愿 | 标题是否让目标读者愿意点开,而非为了新奇制造陌生感? |
| 任务完成度 | 内容是否解决读者此刻要找答案、做选择、避风险或获得共鸣的任务? |
| 理解成本 | 读者能否迅速明白文章在说什么,概念与转折是否超出必要程度? |
| 可信度 | 事实、经验边界、说话位置和不确定性是否交代得自然可信? |
| 活人感 | 是否有真实取舍、具体观察和自然节奏,而非只删除几个 AI 高频词? |
| 信息增量 | 各段是否持续增加事实、动作、区分、推理或后果? |
| 平台原生度 | 标题、段落、语气、长度和交付形态是否像目标平台内容? |
| 记忆度 | 读完后能否留下一个清楚判断或可复述的方法?不强求金句。 |

## 三、失败信号

- 所有题目都被升级成宏大立意、反常识概念或“不是 A,而是 B”。
- 标题变新了,但读者不知道文章与自己有什么关系。
- 同题多跑时只是替换同义词,开头和结构仍相同。
- 为了显得有深度增加材料无法支撑的心理解释、故事细节或权威口吻。
- 基础指南缺少直接答案,情感文章却被数据和专家说明压住。
- 文章看似流畅,但连续段落没有新增信息或行动。

## 四、建议门槛

- 任一结果的“任务完成度”“可信度”“平台原生度”低于 3 分,不进入交付。
- 新版本的平均“点击意愿”或“活人感”提升,不能以“理解成本”明显恶化为代价。
- 同题 3 次结果应至少在标题入口、开头动作、材料组织或结构路径中的两个维度形成实质差异。
- 评审者分歧较大时,保留原文与理由,优先修正规则边界,不把个人偏好写成新的全局硬规则。

## 五、评测记录

每次规则升级记录:版本、模型、平台、种子题、创新档位、增强级别、各项评分、主要问题和是否通过。没有实际盲评数据时,只能写“已建立评测基准”,不能宣称效果已被证明。
slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack."Stars: 198kUpdated Claude CodeCodexVerifiedInstall
ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report.Stars: 194kUpdated Claude CodeCodexVerifiedInstall
ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt.Stars: 194kUpdated Claude CodeCodexVerifiedInstall