wechat-diary

name	wechat-diary
description	Umbrella workflow for turning WeChat public account articles into SecondBrain diary notes. Use when the user provides an mp.weixin.qq.com article and asks to write it as a diary, archive the article's original正文 under StudyLogNew/diary, include downloaded images, and add Codex's understanding inline with the original article content.

WeChat Diary

Overview

Use this skill to coordinate the existing wechat-article-fetcher and secondbrain-diary skills. Do not reimplement article fetching or diary conventions here; load and follow those skills when available, then apply the integration rules below.

The output note is an article archive with commentary, not a summary-only note. The WeChat article's original正文 must be present in the final diary entry unless the fetch itself failed to obtain readable article text.

Workspace Scope

• Final diary edits belong in /home/spreadzhao/workspaces/SecondBrain.
• Fetch raw article Markdown into a temporary path first; copy only selected images and final diary content into the SecondBrain diary tree.
• Do not create raw article archives or resource folders under the spreadconfig skill source directory.

Dependent Skills

• Use wechat-article-fetcher to fetch the article Markdown and local images.
• Use secondbrain-diary to choose the diary path, frontmatter shape, resource directory, and Obsidian link style.
• Use the same external-context and existing-note standards as secondbrain-conversation-diary, adapted to source-driven WeChat article notes.
• If the user explicitly names either dependent skill, still treat this skill as the umbrella workflow that orders them.

Workflow

1. Check repository state before editing and preserve unrelated user changes.
2. Fetch the WeChat article into a temporary Markdown file with images downloaded locally.
3. Read the fetched article before writing. Use the article title, source URL, author, structure, images, and tables as source material.
4. Search existing SecondBrain or Obsidian notes for related private context when the article topic may connect to prior diary entries, projects, decisions, people, concepts, books, articles, technical investigations, or unresolved questions.
5. Collect or refresh relevant external sources when the article involves public facts, technical details, current events, community practice, comparable discussions, official documentation, papers, issues, blogs, videos, news, or social/community references that would make the diary more grounded.
6. Create or update one diary note under StudyLogNew/diary/<year>/; do not create a separate raw article note.
7. Copy article images into StudyLogNew/diary/<year>/resources/ with deterministic names such as wechat_<article-id>_image01.png.
8. Rewrite image references as Obsidian embeds using repository-root paths, for example ![[StudyLogNew/diary/2026/resources/wechat_<id>_image01.png]].
9. Preserve the article's original正文 in the diary note. Keep the original section order, paragraphs, examples, tables, captions, references, and meaningful wording from the fetched article. A paraphrase, outline, or summary is not an acceptable substitute for the article body.
10. Remove only obvious WeChat UI chrome, account promotion blocks, QR codes, star/share prompts, and unrelated recommendation blocks. If unsure whether a paragraph is article content, keep it.
11. Add Codex's understanding near the original paragraphs or sections it explains, usually as > [!note] 我的理解 callouts. Do not put all understanding in a separate second document, and do not let commentary replace the article text.
12. Relate the article to the user's active projects or recurring themes when there is clear context from nearby notes, but keep claims traceable to the article, external sources, or local notes.

Original Article Preservation

Treat source preservation as a hard requirement:

• The final diary must contain the WeChat article正文 itself, not only a synthesized summary, extracted outline, or Codex rewrite.
• Preserve fetched Markdown paragraphs close to the article wording. Cleanup is allowed only for spacing, broken Markdown, obvious duplicated title/author chrome, and local image-link rewriting.
• Keep article headings as real Markdown headings when possible. If the source uses bold section titles, converting them to ## headings is acceptable when it improves readability without changing meaning.
• Keep the author's examples, named people, timelines, numbers, tables, quotes, caveats, and conclusion unless they are part of non-article UI/promotion blocks.
• Insert 我的理解 callouts after the relevant original section. The callouts should explain, connect, or contextualize; they must not replace the original section.
• If the fetcher returns only a verification page, summary, broken extraction, or otherwise insufficient正文, retry with the fetcher options or report the blocker. Do not silently create a summary-only diary.
• If copyright or source-quality concerns require shortening a public-facing response, still write the local diary according to this skill's archive requirement and keep the final user response brief.

Diary Shape

Use the current local date unless the user asks for another date. Prefer a concise English slug that describes the article topic, for example:

StudyLogNew/diary/2026/2026-06-22_loop-engineering-loop-vs-goal.md

Use this structure:

---
title: Human Readable Title
date: YYYY-MM-DD
tags:
  - ai
mtrace:
  - YYYY-MM-DD
star: false
description: One concise sentence
---

参考：

- 原文：[Article Title](https://mp.weixin.qq.com/s/...)

# Article-Or-Topic Title

After the H1, include the article's original正文 as the main body, with local image embeds and 我的理解 callouts interleaved. Sections such as ## 关联笔记, ## 外部参考, and ## 留给自己的判断 belong after the preserved article body unless a source paragraph specifically needs an immediate note.

Choose tags from nearby diary notes when possible. Use Chinese prose for the diary body unless the surrounding note style or source material clearly suggests otherwise.

Keep star: false by default. Set star: true only when the user explicitly asks to star, favorite, highlight, pin, or put the diary in the Star section.

External Context Research

Use external research as part of WeChat diary writing whenever it can improve factual grounding, surface similar discussions, or add useful context for the future note.

• Before making definitive claims about public, technical, current, or disputed matters, search for relevant sources instead of relying only on model memory or the WeChat article.
• Look for concrete references across official documentation, papers or preprints, forums, GitHub issues, blogs, videos, news, and social or community posts.
• Prefer primary or authoritative sources for factual claims. Use forums, issues, videos, and community discussions to capture practical experience, tradeoffs, and disagreement.
• Cite direct links in the diary when external sources shape the writing.
• Keep source use proportional, but for technical articles default to adding at least a small ## 外部参考 or ## 相关资料 section when useful sources are available.
• If no useful source is found, or browsing is unavailable, say that clearly. Do not fabricate links, titles, publication details, or community consensus.
• Clearly distinguish the WeChat article's claims from external context. External sources can support interpretation, but they must not be used to silently rewrite the article author's position.

Existing Note Context

Use the existing SecondBrain vault as private context when it can make the WeChat diary more connected to the user's prior knowledge.

• Search existing notes by topic keywords, project names, tool names, people, dates, decisions, tags, and likely aliases.
• Prefer Obsidian tooling when available. If Obsidian is unavailable, fall back to repository or filesystem search only when the vault path is available and doing so is safe.
• Use related notes to surface continuity, contradictions, prior learning, active projects, and unresolved questions.
• In the diary, cite related notes with Obsidian wikilinks such as [[Note Title]] when the title is known.
• Add a ## 关联笔记, ## 相关笔记, or equivalent section when related notes materially informed the result.
• Do not dump large private note contents into the diary. Quote only the minimal excerpt needed, and summarize the rest.
• Do not modify related notes unless the user explicitly asks to update those notes.

Integration Rules

• Keep article content and Codex understanding in the same diary note.
• The article content means the original WeChat正文. Do not satisfy this rule with a condensed retelling.
• Put 我的理解 callouts immediately after the relevant article section.
• Prefer a few high-signal callouts over a comment after every paragraph.
• Add external-reference and related-note sections when they materially improve the diary; keep them separate from the article body and clearly trace claims to the article, external sources, or local notes.
• Preserve useful article images and captions; skip only decorative or recommendation-only images when they do not support the note.
• Keep Markdown tables readable and do not rewrite the user's or article author's phrasing unless cleanup is necessary.
• When adding local resources, follow repository attachment rules: folder name resources, under the current diary year folder, linked from the repository root.

Quality Pass

Before finishing:

• Check frontmatter fields with rg.
• Check that the diary contains the article's original正文, not just a summary. At minimum, compare the fetched article's main headings and several distinctive body paragraphs against the final note.
• If the final note is mainly paraphrase, outline, or commentary, treat the workflow as failed and revise before reporting completion.
• Check there are no remaining remote WeChat image links or temporary asset paths.
• Check image embed count against copied resource count when images were downloaded.
• Check external links and related-note wikilinks when those sections were added.
• Check git status --short and report only files created or changed for this workflow.