WeChat Official Account Writing System

Purpose

Automatically complete the full process of official account article writing based on the topic provided by the user (optional reference URLs): multi-source retrieval and evidence pool organization, first draft generation, self-review and polishing, reference material organization, with optional automatic illustration and saving to the official account draft box (no submission for publication).

When to Use

User requests "Write a WeChat official account article about XXX"
User requests "Generate a WeChat official account article with the topic XXX"
User needs a complete WeChat official account article creation process
User hopes to "automatically add illustrations after writing" or "send to the official account draft box after writing"

Prerequisites

Before execution, confirm:

The user has provided the article topic (
```
topic
```
)
If there are reference article URLs, they can be provided together (
```
urls
```
)
If
```
topic
```
is purely Chinese and no
```
slug
```
is provided, it is recommended to supplement with an English/pinyin kebab-case directory name; otherwise, the ASCII downgrade scheme will be used

Workflow

Execute in the following order (products are saved to

articles/<slug>/...

by default, facilitating re-running and traceability):

Phase 0: Preflight

Goal: Determine stable and reusable directory and path specifications

Operations:

Calculate
```
slug
```
- If the user provides
```
slug
```
  : use it directly (recommended: English/pinyin kebab-case)
- If
```
topic
```
  contains Latin letters/numbers: convert it to kebab-case
- Otherwise, downgrade to:
```
wechat-article-YYYYMMDD
```

Create directories:

```
articles/<slug>/
```
```
articles/<slug>/sources/
```

Specification: Markdown image references must use relative paths and
```
/
```
separators

Step 1: Material Collection

Goal: Collect materials related to the topic and organize them into a traceable evidence pool

Operations:

If the user provides
```
urls
```
: use
```
webfetch
```
in parallel to obtain content, extract key points, and record the URL and available publication date
If the user does not provide
```
urls
```
: use
```
WebSearch
```
in parallel for multi-source retrieval (recommended coverage: official documents / X(Twitter) / Reddit / technical forums / WeChat official accounts / engineering practices)
- official/authority: official documents, standards/specifications, authoritative media interpretations
- community: X(Twitter), Reddit, forums/discussions
- practice: GitHub issues, engineering blogs, case reviews
- Recommended parallel query templates (combine as needed, try to add years/time ranges to emphasize recent content):
  - {topic} official documentation
    /
    {topic} release notes 2025 2026
  - {topic} site:x.com
    /
    {topic} site:twitter.com
  - {topic} site:reddit.com
    /
    {topic} site:reddit.com/r/<subreddit>
  - {topic} site:github.com issues
    /
    {topic} site:github.com discussions
  - {topic} site:mp.weixin.qq.com
    (WeChat official accounts)
  - {topic} practical review pitfalls 2025 2026
    (Chinese engineering practices)
Merge and deduplicate, grade by credibility (high/medium/low), form an evidence pool, and save it to:
- ```
articles/<slug>/sources/evidence.md
```

Tool Mapping:

"Search" in this process uses:
```
WebSearch
```
"Fetch web page content" in this process uses:
```
webfetch
```

About WebSearch (Implementation Notes):

Priority is given to using the
```
WebSearch
```
tool built into the runtime environment.
If no available
```
WebSearch
```
is present in the current environment: use
```
webfetch
```
to crawl public search results pages (SERP), extract the URL list from the results, then use
```
webfetch
```
in parallel to obtain the main content.

Evidence Pool Entry Format (Each entry must include):

title
url
published_at (if available)
source_type (official/community/practice)
key_takeaways (3-6 key points, preferably directly adaptable to main content materials)
confidence (high/medium/low)

Stop Conditions (Recommended):

8-12 candidate sources, with medium/high >= 5 sources

Common Failure Handling (Newly Added):

X/Reddit login wall or inability to crawl main content:
- Prioritize publicly accessible mirrors/references (second-hand reports must be marked
```
confidence=low/medium
```
  ), or switch to crawling blogs/forum reposts with the same viewpoint;
- If original posts must be cited: only use WebSearch result summaries + other independent sources for corroboration, do not fabricate details.
SERP crawling/parsing failure (when no WebSearch fallback path is available):
- Switch to the "site-specific retrieval" strategy: directly use
```
webfetch
```
  to crawl the official site's search/blog index page or GitHub search page;
- If coverage is still not possible: ask the user for 3-5 key URLs or a specified list of information sources.
Deduplication and credibility:
- At least 2 independent sources to corroborate the same fact;
- Official/primary documents are prioritized as
```
high
```
  ; community discussions without implementation details or cross-validation are marked
```
low
```
  .

Output:

sources_path

(evidence pool path)

Step 2: First Draft Generation

Goal: Generate the first draft of the official account article based on materials

Writing Requirements:

Title: Eye-catching, can use the following techniques
- Number type:
```
5 Methods to Help You...
```
- Question type:
```
Why...?
```
- Comparison type:
```
Differences Between A and B...
```
- Suspense type:
```
Things You Don't Know About...
```
Opening (First 3-5 lines):
- Present pain points or resonate with readers
- Set suspense or raise questions
- Clarify the value of the article
Main Content Structure:
- Use subheadings for segmentation (## or ###)
- 200-300 words per paragraph
- Include cases and data support
- Use lists to enhance readability
Conclusion:
- Summarize core viewpoints
- Guide interaction (likes, "read later", comments)
- Can add golden sentences or calls to action
Word Count Requirement: 1500-2500 words
Traceability Requirement (New Hard Constraint):
- Key facts/data/conclusions must be supported by the evidence pool
- The end of the article must append
```
## References/Sources
```
  (5-10 links, preferably with dates)

Output: First draft in Markdown format, saved to:

articles/<slug>/article.md

Step 3: Intelligent Review

Goal: Review the article from four dimensions to identify issues

Review Dimensions:

Dimension	Check Points	Weight
Logic	Clear arguments, sufficient evidence, reasonable reasoning	30%
Expression	Fluent language, no AI traces, appropriate colloquialism	25%
Data	Accurate data, appropriate cases, standardized citations	25%
Structure	Eye-catching title, clear paragraphs, echo between beginning and end	20%

New Hard Checks:

Traceability: Whether key assertions are supported by the evidence pool
Title Consistency: Whether the title's promise is clearly fulfilled in the main content

Review Operations:

Check logical coherence paragraph by paragraph
Mark AI trace vocabulary (such as "To sum up", "It is not difficult to see", etc.)
Verify the accuracy of data and cases
Check structure and rhythm

Scoring Standards:

90-100 points: Excellent, can be published directly
80-89 points: Good, minor optimizations needed
70-79 points: Qualified, needs revision
Below 70 points: Needs rewriting

Output: Review report (including issues and revision suggestions), recommended to be saved to:

articles/<slug>/sources/review.md

Step 4: Polishing

Goal: Fix issues and improve article quality

Polishing Operations:

Remove AI Traces

Replace vocabulary:

```
综上所述
```
→
```
All in all
```
```
总而言之
```
→
```
To put it simply
```
```
由此可见
```
→
```
So we can see
```
```
不难看出
```
→
```
We can find that
```
```
众所周知
```
→
```
As we all know
```

Avoid overly formal expressions

Enhance Colloquialism
- Appropriately add:
```
Actually
```
  ,
```
To be honest
```
  ,
```
I have to say
```
- Use short sentences, avoid long and complex sentences
- Add transition words to enhance fluency
Optimize Rhythm
- Combine long and short sentences
- Appropriately use exclamations and rhetorical questions
- Control paragraph length (recommended no more than 5 lines)
Fix Review Issues
- Fix items one by one according to the review report
- Supplement missing data or cases
- Adjust unreasonable structures

Output: Final article

Mandatory Requirements (Newly Added):

The final article must retain or supplement
```
## References/Sources
```
It is recommended to keep 5-10 references, grouped by
```
official
```
/
```
community
```
/
```
practice
```
is better

Step 5: Saving and Output

Operations:

Create output directory (if not exists):
- ```
articles/<slug>/
```
Save article:
- ```
articles/<slug>/article.md
```
Output execution summary:
- ```
article_path
```
- ```
sources_path
```
- Word count
- Review score

Step 6: Automatic Illustration (Optional, call zhy-article-illustrator)

Trigger Condition:

with_illustrations=true

Goal: Generate high-quality illustrations with a unified style for the article, and produce an illustrated version of the article

Default Strategy:

```
article_path
```
: Use
```
articles/<slug>/article.md
```
```
slug
```
: Reuse the current article slug
```
density
```
:
```
illustration_density
```
(default balanced)
```
upload
```
:
```
illustration_upload
```
(default false)
```
aspect_ratio
```
:
```
illustration_aspect_ratio
```
(default 16:9)

prompt_profile

illustration_prompt_profile

(default nano-banana)

```
text_language
```
:
```
illustration_text_language
```
(default zh-CN)

english_terms_whitelist

illustration_english_terms_whitelist

(default empty)

image_provider

illustration_image_provider

(default xiaomi)

```
image_model
```
:
```
illustration_image_model
```
(default gemini-3.1-flash-image-preview)
```
image_size
```
:
```
illustration_image_size
```
(default 1K)
```
image_base_url
```
:
```
illustration_image_base_url
```
(default Xiaomi API address, also supports Gemini native proxy)

Execution Method:

Prioritize calling the one-click process script of

zhy-article-illustrator

bash

node <zhy-article-illustrator>/scripts/illustrate-article.ts \
  --article articles/<slug>/article.md \
  --slug <slug> \
  --density <illustration_density> \
  --aspect-ratio <illustration_aspect_ratio> \
  --prompt-profile <illustration_prompt_profile> \
  --text-language <illustration_text_language> \
  --image-provider <illustration_image_provider> \
  --image-model <illustration_image_model> \
  [--image-size <illustration_image_size>] \
  [--image-base-url <illustration_image_base_url>] \
  [--upload]

illustration_english_terms_whitelist

is not empty, append

--term <value>

for each term, for example:

bash

--term Playwright --term Chromium --term Firefox --term WebKit

Default to the new illustration strategy:
- First generate the article-level
```
visual-bible.md
```
- Then generate
```
outline.md
```
  and
```
prompts/
```
- Default text in images is Simplified Chinese, only whitelisted terms retain English
- All images in the same article share a unified style system

When integrating writing skills, follow the following field mapping:

```
article_path -> --article
```
```
slug -> --slug
```
```
illustration_density -> --density
```

illustration_aspect_ratio -> --aspect-ratio

illustration_prompt_profile -> --prompt-profile

illustration_text_language -> --text-language

illustration_image_provider -> --image-provider

illustration_image_model -> --image-model

```
illustration_image_size -> --image-size
```

illustration_image_base_url -> --image-base-url

```
illustration_upload=true -> --upload
```

Output:

illustrated_article_path

articles/<slug>/article.illustrated.md

illustrations_dir

articles/<slug>/illustrations/<slug>/

articles/<slug>/illustrations/<slug>/visual-bible.md

articles/<slug>/illustrations/<slug>/outline.md

articles/<slug>/illustrations/<slug>/prompts/

Failure Handling: Retry once for single image failure; if still failed, record and continue, finally output the failure list. If some images fail,

article.illustrated.md

should still be retained, and image placeholder comments inserted.

Step 7: HTML Theme Style Output (Optional, call zhy-markdown2wechat)

Trigger Condition:

with_html_theme=true

Goal: Use the

zhy-markdown2wechat

skill to convert Markdown to HTML with WeChat inline styles

Operations:

Select input file:

with_illustrations=true

and

articles/<slug>/article.illustrated.md

exists, use this file

Otherwise, use
```
articles/<slug>/article.md
```

Record the selected Markdown file as
```
<input_markdown>
```
Call the
```
zhy-markdown2wechat
```
skill and execute the conversion script:
bash
```
node <zhy-markdown2wechat>/scripts/convert.js \
 <input_markdown> \
 <zhy-markdown2wechat>/resources/themes/default.css \
 articles/<slug>/article.zhy.html
```
- The script has zero dependencies (pure Node.js), no need for
```
npm install
```
  , automatically processes in a temporary directory and cleans up afterwards
- Output includes a
```
<section id="MdWechat">
```
  container and complete inline CSS styles
- To change the theme, replace the second parameter with other theme files under
```
resources/themes/
```
  (
```
apple.css
```
  /
```
blue.css
```
  /
```
dark.css
```
  /
```
green.css
```
  /
```
notion.css
```
  /
```
vibrant.css
```
  )
- ```
<zhy-markdown2wechat>
```
  refers to the installation directory of this skill in the current environment, use the actual path during runtime

Input file examples:

If illustrated article exists:

<input_markdown>=articles/<slug>/article.illustrated.md

If illustrated article does not exist:

<input_markdown>=articles/<slug>/article.md

Output:

html_article_path

(

articles/<slug>/article.zhy.html

)

Failure Handling: Record the error and note the reason in the execution summary, skip this step and continue with subsequent processes

Step 8: Save to Official Account Draft Box (Optional, call zhy-wechat-publish)

Trigger Condition:

post_to_wechat=true

Default Behavior: Save to the draft box via the official WeChat API, no final publication submission

Prerequisites:

```
WECHAT_APP_ID
```
and
```
WECHAT_APP_SECRET
```
have been configured in the
```
.env
```
file under the
```
zhy-wechat-publish
```
skill directory
The public IP of the running machine has been added to the IP whitelist in the WeChat official account backend
To automatically generate a cover image, the image generation environment dependent on the publishing skill must also be available (provided by
```
zhy-article-illustrator
```
)

Calling Method:

The main content must be an HTML file with inline styles
Prioritize using
```
article.zhy.html
```
generated in Step 7; if it does not exist, skip this step (or first complete Step 7)

The recommended stable entry is to directly call

wechat_draft.js

bash

node <zhy-wechat-publish>/scripts/wechat_draft.js \
  --title "Article Title" \
  --file "articles/<slug>/article.zhy.html" \
  [--author "Author"] \
  [--digest "Abstract"] \
  [--thumb "Cover media_id"] \
  [--source-url "Original Link"] \
  [--need-open-comment "1"] \
  [--only-fans-can-comment "1"]

To automatically generate a cover image and publish, you can also call:

bash

node <zhy-wechat-publish>/scripts/publish_with_cover.js \
  --article "articles/<slug>/article.md" \
  --html "articles/<slug>/article.zhy.html" \
  [--title "Article Title"] \
  [--author "Author"] \
  [--source-url "Original Link"] \
  [--need-open-comment "1"] \
  [--only-fans-can-comment "1"]

```
<zhy-wechat-publish>
```
refers to the installation directory of this skill in the current environment, use the actual path during runtime

When

--thumb

is not provided in

wechat_draft.js

, it will automatically read

WECHAT_DEFAULT_THUMB_MEDIA_ID

from

.env

```
publish_with_cover.js
```
will automatically extract the title/abstract from the article, generate a 16:9 cover image, upload the cover image, and use the returned
```
media_id
```
as
```
thumb_media_id
```
The publishing script will automatically expand
```
var(--xxx)
```
style variables in the HTML before uploading to avoid loss of colors and border styles in the WeChat draft box
The publishing script will automatically upload images in the main content to the WeChat main content image interface before uploading, and replace
```
<img src>
```
with the image URL returned by WeChat
The publishing script will automatically downgrade the native list structure to "ordinary paragraphs + dots/numbers" before uploading to resolve the list parsing issue when re-entering edit mode in the WeChat draft box

Notes:

The script has zero dependencies (pure Node.js >= 16), no need for
```
npm install
```
When using
```
publish_with_cover.js
```
,
```
bun
```
must be available locally, as the cover image generation reuses existing image generation scripts
After saving the draft, it will not be automatically submitted for publication; manual confirmation in the official account backend is required
The title length cannot exceed 64 characters
If no available image generation configuration exists in the current environment, prioritize using
```
wechat_draft.js
```
to directly upload HTML to avoid failure in the automatic cover image step

Success Criteria: Output

Upload draft successfully! Draft MEDIA_ID: xxx

Failure Troubleshooting List:

Error Message	Cause and Handling
`[40013] invalid appid`	Incorrect AppID, check `.env` under the publishing skill directory
`[40164] invalid ip`	Current IP not in whitelist, add the IP in the error message to the official account backend
`[40007] invalid media_id`	Invalid cover image ID, re-upload using `upload_image.js` to obtain a new one
`Missing Xiaomi/Gemini/OpenAI API Key`	Image generation environment dependent on automatic cover image generation not configured, check `.env` related to `zhy-article-illustrator`
`article.zhy.html` does not exist	Step 7 not executed or failed, check `with_html_theme=true`
Title too long	Control title <= 64 characters

Data Flow

User Input (topic, urls?, slug?, search_count?, time_range_days?, ...)
         ↓
Preflight (Determine slug and directory)
         ↓
Material Collection (WebSearch + webfetch → evidence.md)
         ↓
First Draft Generation (article.md)
         ↓
Intelligent Review (including traceability/title consistency)
         ↓
Polishing (Mandatory References)
         ↓
Automatic Illustration (article.illustrated.md + illustrations/)
         ↓
HTML Theme Style Output (zhy-markdown2wechat → article.zhy.html)
         ↓
Save to Draft Box (No submission for publication)

Error Handling

Exception	Handling Method
No search results	Prompt user to provide more information or reference URLs
Reference article inaccessible	Skip this URL and continue processing other materials
Low quality of first draft	Regenerate or prompt user to provide more materials
Review score <70	Suggest user check if the topic is appropriate
Illustration failure	Output failure list; can choose to supplement images before publishing
HTML conversion failure	Record error and skip this step (Step 7), continue with subsequent processes
Failure to save to draft box	Output troubleshooting list (AppID/IP whitelist/cover media_id/title length)

Example Usage

Input:

topic: "How to Improve Work Efficiency"
urls: ["https://mp.weixin.qq.com/xxx"]
search_count: 5
with_illustrations: true
with_html_theme: true
post_to_wechat: true

Execution Process:

Search for articles related to "How to Improve Work Efficiency"
Obtain content of the reference article provided by the user
Generate first draft (about 1500-2500 words)
Review score: 85 points
Polish and optimize, then save

Output:

article_path: articles/how-to-improve-work-efficiency/article.md
sources_path: articles/how-to-improve-work-efficiency/sources/evidence.md
illustrated_article_path: articles/how-to-improve-work-efficiency/article.illustrated.md
illustrations_dir: articles/how-to-improve-work-efficiency/illustrations/how-to-improve-work-efficiency/
html_article_path: articles/how-to-improve-work-efficiency/article.zhy.html
word_count: 2150
review_score: 92
wechat_draft_status: success

Notes

The article style should conform to the official account's tone: relaxed, useful, and resonant
Avoid sensitive content and excessive marketing
Maintain originality, do not directly copy material content

zhy-wechat-writing

NPX Install

Tags

SKILL.md Content (Chinese)

WeChat Official Account Writing System

Purpose

When to Use

Prerequisites

Workflow

Phase 0: Preflight

Step 1: Material Collection

Step 2: First Draft Generation

Step 3: Intelligent Review

Step 4: Polishing

Step 5: Saving and Output

Step 6: Automatic Illustration (Optional, call zhy-article-illustrator)

Step 7: HTML Theme Style Output (Optional, call zhy-markdown2wechat)

Step 8: Save to Official Account Draft Box (Optional, call zhy-wechat-publish)

Data Flow

Error Handling

Example Usage

Notes