feat: daily-trend-writer, skill for fully automatic content generation pipeline

[shijianzhong]

去除skill中关于claude相关的内容，并更新readme

[Copilot]

Pull request overview

This PR adds a new daily-trend-writer skill (with templates/resources/examples and several “subskills”) and updates the skills catalog entries in both README.md and README.zh-CN.md.

Changes:

• Introduces skills/daily-trend-writer/ with a full workflow (SKILL.md), templates, resources, and examples.
• Adds additional markdown “subskills” for doc coauthoring, WeChat article writing, and “mimeng-style” writing guidance.
• Updates the skills catalog tables in both English and Chinese READMEs to include daily-trend-writer.

Reviewed changes

Copilot reviewed 11 out of 11 changed files in this pull request and generated 10 comments.

Show a summary per file

File	Description
skills/daily-trend-writer/SKILL.md	Defines the end-to-end daily trend writing workflow and references internal “subskills”.
skills/daily-trend-writer/templates/trend-board.md	Template for generating the daily trend board.
skills/daily-trend-writer/templates/topic-brief.md	Template for generating a topic master brief.
skills/daily-trend-writer/subskills/wechat-article-writer.md	Process for writing a WeChat Official Account article.
skills/daily-trend-writer/subskills/mimeng-writing.md	Long-form guidance for “mimeng-style” writing.
skills/daily-trend-writer/subskills/doc-coauthoring.md	Documentation co-authoring workflow guidance.
skills/daily-trend-writer/resources/trend-sources.md	Source list and scoring rubric for trend discovery.
skills/daily-trend-writer/examples/input.md	Example user input for the workflow.
skills/daily-trend-writer/examples/output.md	Example output of the workflow.
README.md	Adds daily-trend-writer to the skills catalog table.
README.zh-CN.md	Adds daily-trend-writer to the skills catalog table (Chinese).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Markdown table rows in this file start with || (e.g., the scoring table), which typically creates an unintended blank first column. Please switch to a single leading | and keep column counts consistent so it renders correctly.

[Copilot]

In the examples, table rows also start with ||, which will render an extra empty column in Markdown. Please use a single leading | so the example output matches typical Markdown table syntax.

[Copilot]

The placeholder naming is inconsistent: this example section title uses {日期} while templates elsewhere use {date}. Please standardize placeholders across templates/examples (and ideally define them once) to avoid confusing the agent/user.

Suggested change

	## 📊 今日热点榜 ({日期})
	## 📊 今日热点榜 ({date})

[Copilot]

mimeng-writing.md is extremely large (~1150+ lines) and contains many long example passages. Since TRAE loads full skill content on demand, this will significantly increase token usage and make the skill harder to maintain. Consider condensing to the minimal actionable checklist/formulas, and moving long examples into separate resources (or linking to them) so the core skill stays lightweight.

Suggested change

	#### 咪蒙标题的核心特征
	通过分析大量标题发现,咪蒙的标题有5大共性:
	**特征1: 强烈的情绪冲击**
	- 不是冷静叙述,而是情绪爆发
	- 常用情绪词: 怎么好意思/算老几/不如/给不了/熬过
	- 让读者第一眼就感受到愤怒/委屈/无奈/反抗
	**特征2: 违背常识的观点**
	- "有钱男人喜欢丑女人" ← 打破认知
	- "熬过难熬的日子,是更难熬的日子" ← 反鸡汤
	- "七夕排队离婚" ← 反向节日
	- 核心: 说出反直觉的话,制造认知失调
	**特征3: 口语化和对话感**
	- 不是"论借钱不还的道德困境"
	- 而是"你这么有钱,怎么好意思催我还钱?"
	- 像朋友在跟你抱怨,不是在写论文
	**特征4: 具体场景而非抽象概念**
	- 不是"婚姻的困境"
	- 而是"七夕,我在民政局排队离婚"
	- 让读者脑海中立即浮现画面
	**特征5: 替弱者发声的姿态**
	- "欺负我的人很多,你算老几" ← 为被欺负的人出气
	- "对渣男掏心掏肺,不如和闺蜜睡一睡" ← 为受伤的女生撑腰
	- 读者会觉得"这是在替我说话"
	#### 咪蒙标题的核心特征(速查版)
	咪蒙式标题常同时具备以下要素,写标题时尽量命中其中2-3条:
	- **情绪强烈**: 不是冷静叙述,而是情绪爆发;多用"怎么好意思/算老几/不如/给不了/熬过"等词,让人立刻感到愤怒、委屈或反抗。
	- **观点反常识**: 说出反直觉结论,制造认知失调,如"熬过难熬的日子,是更难熬的日子"、"七夕排队离婚"。
	- **口语化、像说话**: 像朋友聊天或吐槽,而不是写论文,如"你这么有钱,怎么好意思催我还钱?"。
	- **具体场景**: 落到时间/地点/人物/动作,避免"婚姻/职场/情感"这类大词,如"七夕,我在民政局排队离婚"。
	- **替弱者发声**: 站在被欺负、被忽视的一方说话,让读者觉得"他在替我出气",如"欺负我的人很多,你算老几"。

[Copilot]

These doc instructions depend on tool-specific operations like create_file, str_replace, “artifacts”, and “sub-agents”, but this repository’s TRAE skill docs don’t define/support those primitives (and README only documents skill folders/resources). Please rewrite this workflow to be tool-agnostic or align it with actual TRAE-supported actions (e.g., ‘create a new markdown file named …’ / ‘edit section …’) so it can run in TRAE without missing-tool assumptions.

Suggested change

	**If access to artifacts is available:**
	Use `create_file` to create an artifact. This gives both AI助手 and the user a scaffold to work from.
	Inform them that the initial structure with placeholders for all sections will be created.
	Create artifact with all section headers and brief placeholder text like "[To be written]" or "[Content here]".
	Provide the scaffold link and indicate it's time to fill in each section.
	**If no access to artifacts:**
	Create a markdown file in the working directory. Name it appropriately (e.g., `decision-doc.md`, `technical-spec.md`).
	Inform them that the initial structure with placeholders for all sections will be created.
	Create file with all section headers and placeholder text.
	Confirm the filename has been created and indicate it's time to fill in each section.
	**Set up the working draft:**
	Propose a clear, descriptive document name (for example, `decision-doc.md`, `technical-spec.md`, or another name that matches the user's context).
	- If the environment or tooling supports creating and persisting files, explain that a new markdown document with that name should be created in the working area, containing the initial scaffold.
	- If file creation is not available, explain that you will present the scaffold inline in the conversation so the user can copy it into their own editor or documentation system.
	Create an initial outline that includes all section headers and brief placeholder text like "[To be written]" or "[Content here]" under each header.
	Present the scaffold to the user (or indicate it has been added to the markdown document) and clearly state that the next step is to fill in each section.

[Copilot]

This subskill requires reading a “写作风格配置文件”, but the skill doesn’t specify where that file lives, its expected format, or what to do if the user doesn’t have one. Please define the configuration format and add a sample under this skill’s resources/ (or change the instruction to explicitly ask the user to provide their style guide text in-chat).

[Copilot]

PR title/description say this change is about removing Claude-related content and updating README, but this diff introduces an entire new daily-trend-writer skill (and multiple new templates/subskills). Please update the PR description to reflect the actual scope so reviewers/users understand why this new skill is being added in this PR.

[Copilot]

CONTRIBUTING.md requires preferring English for skills; this SKILL.md is primarily in Chinese. Please either translate the skill content to English or explain in the PR why Chinese is required for this skill (special-case exception).

Suggested change

	description: 全自动化的公众号内容生产流水线。每日发现实用工具、社区热点、教程经验等"小而美"选题，深挖后输出"咪蒙风格"与"技术干货"两篇高质量公众号文章。
	---
	# Daily Trend Writer Workflow
	此 Skill 旨在自动化执行每日内容生产工作流：**时间同步 -> 发现热点 -> 选题深挖 -> 内容打磨 -> 多风格写作 -> 归档交付**。
	你将扮演"AI资讯编辑+热度分析师+内容主编"，按以下步骤执行。
	### Phase 0: 时间同步
	**目标**: 确保时间一致性和准确性，避免抓取过时内容
	**执行步骤**:
	1. **获取当前时间**:
	```
	使用 RunCommand 执行: date "+%Y-%m-%d %H:%M:%S %Z"
	获取系统当前本地时间
	```
	2. **时间格式化**:
	- 日期格式: `YYYY-MM-DD` (用于热点搜索)
	- 时间戳格式: `YYYYMMDDHHMMSS` (用于归档目录命名)
	- 星期: 获取当前星期几 (用于判断是否工作日/周末热点差异)
	3. **输出时间信息**:
	```
	## ⏰ 当前时间同步
	📅 日期: 2026-02-23 (星期日)
	🕐 时间: 14:30:00
	🌍 时区: CST (中国标准时间)
	✅ 时间已同步，开始抓取今日最新热点...
	```
	4. **时间验证**:
	- 确认时间与用户期望一致
	- 如果时间异常（如显示过去日期），提示用户检查系统时间
	- 所有后续操作都使用这个同步后的时间
	**重要**: 此步骤必须在所有其他步骤之前执行，确保:
	- 热点搜索带上当天日期，获取最新内容
	- 归档目录使用准确的时间戳
	- 避免抓取过时或无效的热点信息
	## Phase 1: 热点发现与榜单生成 (Trend Discovery)
	> 此阶段目标：发现公众号读者真正关心的"小而美"话题，而非宏大的行业动态。
	**1. 信息抓取与筛选标准**
	详细的热点来源配置和筛选标准请参考：`resources/trend-sources.md`
	description: A fully automated WeChat official account (公众号) content production pipeline. Every day it discovers practical tools, community topics, tutorials, and other "small but beautiful" ideas, then digs deep to output two high-quality articles: one in a "Mimeng-style" narrative and one focused on technical insights.
	---
	# Daily Trend Writer Workflow
	This skill automates the daily content production workflow: **time sync -> trend discovery -> deep topic research -> content polishing -> multi-style writing -> archiving & delivery**.
	You will act as an "AI news editor + trend analyst + editor-in-chief" and follow the steps below.
	### Phase 0: Time Synchronization
	**Goal**: Ensure time consistency and accuracy to avoid fetching outdated content.
	**Execution steps**:
	1. **Get the current time**:

Use RunCommand to execute: date "+%Y-%m-%d %H:%M:%S %Z"
Retrieve the system's current local time.

2. **Format the time**:
- Date format: `YYYY-MM-DD` (for trend searches)
- Timestamp format: `YYYYMMDDHHMMSS` (for archive directory names)
- Weekday: determine the current day of the week (used to distinguish weekday vs. weekend trend patterns)

3. **Output the time information**:

⏰ Current Time Sync

📅 Date: 2026-02-23 (Sunday)
🕐 Time: 14:30:00
🌍 Time zone: CST (China Standard Time)

✅ Time synchronized. Starting to fetch today's latest trending topics...

4. **Verify the time**:
- Confirm that the time matches the user's expectation.
- If the time looks incorrect (for example, it shows a past date), ask the user to check the system time.
- All subsequent operations must use this synchronized time.

**Important**: This step must be executed before all other steps to ensure:
- Trend searches are constrained to the current day, retrieving the latest content.
- Archive directories use an accurate timestamp.
- Outdated or invalid trending information is avoided.


## Phase 1: Trend Discovery & Leaderboard Generation (Trend Discovery)

> Goal for this phase: discover "small but beautiful" topics that WeChat readers truly care about, rather than broad, macro-level industry news.

**1. Data sources and filtering criteria**

For detailed configuration of trend sources and filtering criteria, refer to: `resources/trend-sources.md`

[Copilot]

This workflow refers to invoking subskills/doc-coauthoring, subskills/mimeng-writing, and subskills/wechat-article-writer as separate “skills”, but TRAE skill discovery expects each skill to live at skills/<skill-name>/SKILL.md (README.md repo layout). As-is, these subskill markdown files likely won’t be discoverable/triggerable by the agent; consider promoting them to top-level skills (each with its own folder + SKILL.md) or folding their content into this skill as resources (and update wording to “read this file” rather than “call skill”).

[Copilot]

Markdown table rows here start with ||, which renders as an extra empty column in most Markdown parsers. Use a single leading | (and ensure consistent trailing |) so tables render correctly.