audio-to-wechat-article（音频转微信文章）

Audio to WeChat Article

Use this skill to convert raw spoken or written material into a structured WeChat article workflow.

Default outcome:

• transcript or cleaned source text
• article brief
• article JSON
• WeChat-ready markdown
• optional image placement plan
• optional QR code CTA block
• optional WeChat draft publishing

Supported inputs

• audio file (.m4a, .mp3, .ogg, .wav, .mp4, .mov)
• transcript text
• meeting minutes / notes
• optional screenshots or supporting images
• optional QR code image for final article or poster

V5.1 end-to-end workflow

Route A: audio → transcript → article → WeChat draft

1. Transcribe audio into text
2. Build article brief from transcript
3. Decide article mode
4. Draft article JSON
5. Add screenshots / QR placement notes
6. Compose WeChat markdown
7. Optionally publish to WeChat draft box

Route B: text / notes → article → WeChat draft

1. Clean notes into source text
2. Build article brief
3. Draft article JSON
4. Add screenshots / QR placement notes
5. Compose WeChat markdown
6. Optionally publish to WeChat draft box

Core workflow rules

1. Normalize the source

- If the user provides audio, transcribe it first. - If the user provides text, clean it into readable source notes. - Do not fabricate unclear sections; mark uncertain parts.

1. Build an article brief

Use the bundled brief script for a deterministic first pass.

1. Choose article mode

- Insight article: 观点 / 认知 / 判断 - Case article: 复盘 / how we did it - Practical article: 教程 / 流程 / 方法 - Boss-facing article: 管理 / 商业化 / 落地

1. Write the article JSON

Always produce: - title - summary - body - optional cover direction - optional QR CTA sentence

1. Handle images and QR code

- Place screenshots only where they support understanding. - Place QR code near the end by default. - Add one CTA sentence above the QR code.

1. Prepare WeChat-ready markdown

Save markdown with frontmatter: - title - author - summary - coverImage (if available)

1. Optional publish step

If the user asks to publish, hand off the final markdown to baoyu-post-to-wechat.

Writing rules

• Prefer short, strong paragraphs.
• Cut repetitive spoken filler aggressively.
• Keep spoken authenticity, but remove rambling.
• Use subheads to create clear reading rhythm.
• For WeChat, prioritize readability over literal transcript fidelity.
• Keep one article centered on one core message.

Image handling

When the user provides images:

• do not dump images randomly into the article
• decide whether each image is for:

- proof / screenshot evidence - emotional pacing - explanatory diagram - QR code CTA

• place images only where they improve comprehension or conversion

When the user provides a QR code:

• default placement is near the end
• add one sentence CTA above it
• avoid placing QR code too early unless explicitly requested

Recommended output structure

1. 核心观点

• 一句话说清这篇到底讲什么

2. 文章标题候选

• 标题 1
• 标题 2
• 标题 3

3. 正文大纲

• 开头钩子
• 3-5 个主体部分
• 收尾

4. 公众号成稿

Provide full markdown-ready article.

5. 配图建议

• cover direction
• inline image placement notes
• QR placement note (if applicable)

Bundled scripts

• scripts/build_article_brief.py: convert transcript/minutes text into a compact article brief JSON
• scripts/draft_article_json.py: turn source text into a simple article JSON draft
• scripts/compose_wechat_markdown.py: convert article JSON + metadata into WeChat-ready markdown
• scripts/audio_to_article_pipeline.py: end-to-end wrapper from source text/audio metadata to markdown handoff plan

References

• Read references/workflow.md for the full end-to-end process from audio/text to WeChat article.
• Read references/style-guide.md for how to turn spoken content into tighter公众号文章表达.
• Read references/publish-handoff.md for how this skill should hand off to the WeChat publishing workflow.
• Read references/transcription-handoff.md for how to integrate with meeting-minutes-whisper.