Wechat2RSS Feishu
Use
Use this skill for the local Feishu Base workflow around the public Wechat2RSS service:
- initialize a Feishu Wiki child Base for this workflow
- synchronize the official Wechat2RSS account list into the
公众号table - scan enabled accounts and insert newly discovered articles into the
文章table - read stored article data for downstream analysis
This skill assumes lark-cli is already configured by the user. If
lark-cli or the required profile is missing, guide the user to configure
lark-cli; do not create env/json config files and do not document lark-cli
setup internals here.
Do not search local files, shell history, browser tabs, prior chats, or other
places for a missing Wiki URL, Base token, or table ID. If a required runtime
input is not provided, stop and ask the user for it. For initialization, ask
for the parent Wiki URL. For sync or scan, ask for the base_token and table
IDs. If the user cannot provide them, explain which value is missing and guide
them to find or create the target Base.
Use this short user-facing wording when the target data table is unknown:
我需要知道 Wechat2RSS 数据表放在哪里。已经创建过的话,请给我对应的飞书多维表格链接,或初始化输出的 token 和表 ID;还没创建的话,请给我一个飞书 Wiki 父页面链接,我会在下面新建。这是为了定位
公众号表和文章表,后续同步列表、扫描文章、查询文章都要用到它们。
Data Model
Read references/data-model.md before changing tables, fields, deduplication keys, or analysis assumptions.
Scripts
Run scripts with Node.js from this skill directory. Always pass runtime inputs as command-line arguments; do not create persistent config files.
Initialize a Base under a Wiki node:
node scripts/init-base.mjs \
--profile oomol \
--wiki-url "https://oomol-studio.feishu.cn/wiki/..." \
--title "Wechat2RSS"
Synchronize the official account list:
node scripts/sync-official-list.mjs \
--profile oomol \
--base-token "<base_token>" \
--accounts-table "<table_id_or_name>"
Scan enabled feeds and archive new articles:
node scripts/scan-enabled-feeds.mjs \
--profile oomol \
--base-token "<base_token>" \
--accounts-table "<table_id_or_name>" \
--articles-table "<table_id_or_name>"
Read references/workflows.md for execution modes, required arguments, expected output, and failure handling.
Analysis
For analysis requests, read existing 文章 records and use the data model as
the contract. Do not alter Base schema unless the user explicitly asks.
Remember that 正文 is plain text and may be truncated; check 正文是否截断
and 正文长度 when precision matters.
When presenting article references to the user, always render each article as a
Markdown hyperlink using its title and URL, for example [标题](URL). Do not
show a referenced article as plain title text when a URL is available.