WeRead Skills - Setup Guide¶

Source: skills.sh (1 skill, ~41.7K combined installs) GitHub: tencent/wechatreading (225⭐, 14 forks, pushed Jul 1, 2026) Category: Reading & Personal Knowledge First Seen: May 22, 2026 on skills.sh (surfaced in the Sep 11, 2026 sweep) Quality Tier: 🟡 Trusted (Gen Agent Trust Hub Pass / Socket Pass / Snyk Warn)

The official WeChat Reading (微信读书) skill from Tencent - a single Agent Skills-format package that gives an agent read access to one of China's largest digital reading platforms. Search the book catalog, inspect a user's bookshelf, pull notes and highlights, read public reviews, and produce reading statistics. Published under the first-party tencent org, which carries vendor authority on its own; at 41.7K installs it also clears the 20K bar outright.

Installation¶

npx skills add tencent/wechatreading

The README documents the global variant:

npx skills add Tencent/WeChatReading -g

For Hermes Agent, copy the skill folder from a clone:

git clone https://github.com/tencent/wechatreading.git
cp -r wechatreading/skills ~/.hermes/skills/weread-skills

Prerequisites¶

Requirement	Details
WEREAD_API_KEY	Get yours at weread.qq.com/r/weread-skills, then `export WEREAD_API_KEY=wrk-xxxxxxxx`
WeChat Reading account	The API key is bound to a user identity - all user-scoped endpoints (bookshelf, notes, stats) automatically inject it
Chinese-language tolerance	README and SKILL.md are written in Chinese; the agent must be able to follow original-language instructions

Skill Roster¶

Skill	Installs	What It Does
`weread-skills`	41.7K	Umbrella skill for WeChat Reading: book search (e-book / audiobook / web-novel / author / full-text), book details and chapter index, bookshelf management, notes and highlights export, reading statistics (hours, days, rankings, preference analysis), public reviews, and discovery/recommendations

Reference docs shipped alongside the SKILL.md: search.md, book.md, shelf.md, notes.md, readdata.md, review.md, discover.md, profile.md.

Quick Start¶

Install the skill and export WEREAD_API_KEY
Ask in natural language - "帮我搜一下三体" (search for The Three-Body Problem), "看看我的书架" (show my bookshelf), "导出我在这本书里的划线" (export my highlights from this book)
The skill resolves the request to the corresponding WeChat Reading endpoint and returns structured results

CorpusIQ Use Cases¶

Use Case	How
China-market research	Full-text book search surfaces how Chinese business literature covers a topic - positioning and content research without a China phone number
Operator reading systems	Notes and highlights export feeds personal-knowledge extraction for founders who track their reading in WeRead
Review intelligence	Public review browsing gives sentiment and reception data for books, comparable to product-review mining
Reference architecture	A first-party Tencent skill with a user-bound API key is a clean example of identity-scoped agent integrations

Limitations / Verification¶

Snyk Warn: the skills.sh audit renders Gen Agent Trust Hub Pass, Socket Pass, Snyk Warn. The SKILL.md plus 8 reference files are short API-pattern documents - manual review takes minutes. Review before relying on it in production.
Chinese-language content: all instruction text ships untranslated; agents must handle original-language SKILL.md content.
User identity bound: bookshelf, notes, and stats endpoints always act as the key's owner - no shared/team scope.
Platform scope: WeChat Reading is a China-centric service; utility outside Chinese-language reading is limited.

# Verify skill installed
ls ~/.hermes/skills/weread-skills/SKILL.md

Security¶

Audit	Verdict
Gen Agent Trust Hub	Pass
Socket	Pass
Snyk	Warn

← Skills Catalog | Marketplace → Powered by CorpusIQ