feat: add WeRead adapter with 7 commands

[Astro-Han]

Summary

Closes #82. Adds a WeRead (微信读书) adapter with 7 commands covering search, rankings, book details, bookshelf, and note export.

Public commands (no login required, Node.js fetch):

• weread search <keyword> — search books
• weread ranking [category] — book rankings (all / rising / category ID)

Private commands (cookie auth via browser):

• weread book <bookId> — book details
• weread shelf — personal bookshelf
• weread notebooks — list books with highlights/notes
• weread highlights <bookId> — underlines in a book
• weread notes <bookId> — personal notes/thoughts on a book

Architecture

• Shared helper src/clis/weread/utils.ts with two fetch modes:
  • fetchWebApi — Node.js direct fetch for public endpoints (weread.qq.com/web/*)
  • fetchWithPage — browser page.evaluate for private endpoints (i.weread.qq.com/*)
• Each command file is a thin 20-30 line wrapper using cli() registration
• Follows existing patterns from bilibili (browser fetch) and xiaoyuzhou (public fetch + utils in subdirectory)

Verified API endpoints

Endpoint	Status
weread.qq.com/web/search/global	✅ Returns JSON
weread.qq.com/web/bookListInCategory/{cat}	✅ Returns JSON
i.weread.qq.com/*	Requires cookie auth (errcode -2010 without)

Note: Community docs claim book/info and bestbookmarks are public, but our testing confirmed they require cookies.

Test plan

• Unit tests for formatDate, fetchWebApi, fetchWithPage (15 tests)
• E2E tests for weread search and weread ranking in public-commands.test.ts
• Typecheck passes
• Build passes (201 manifest entries)
• Manual smoke test: weread search python, weread ranking all
• Private commands need manual testing with a logged-in Chrome session
• shelf.progress, book.rating, notebooks.noteCount field mappings are from community docs and marked with TODO for runtime verification