feat(weibo): add statuses command to fetch user timeline

[ZepPellN]

Summary

Adds opencli weibo statuses <id> command to fetch a specific user's Weibo timeline.

Currently opencli supports:

• weibo user <id> — user profile
• weibo post <id> — single post detail
• weibo feed — personal for-you / following timeline

But there is no way to list a specific user's statuses. This PR fills that gap.

What Changed

• New adapter: clis/weibo/statuses.js
• Uses /ajax/statuses/mymblog?uid=<uid>&page=<page>&feature=0 API (discovered via browser network capture)
• Resolves screen_name → uid automatically if a non-numeric ID is provided
• Returns rich metadata: id, mblogid, text, isLongText, created_at, reposts, comments, likes, pic_num, url, plus retweeted and page_info when available

Usage

# By UID
opencli weibo statuses 1233486457 --limit 5

# By screen name
opencli weibo statuses "高飞" --limit 10 --page 2

# JSON output
opencli weibo statuses 1233486457 --limit 5 -f json

Test Plan

• opencli validate passes
• E2E tested against real Weibo user pages (authenticated cookie session)
• Both numeric UID and screen_name resolution work
• Long text flag, retweeted content, and page_info are correctly extracted
• Pagination (--page) works

Notes

• Requires cookie authentication (same as other weibo commands)
• text is capped at 500 chars in list view; use weibo post <mblogid> for full long-text expansion

[jackwener]

Thanks for the PR. I reviewed the current head (40f1ae47) against current main; this needs changes before content review can continue.

Required changes:

1. Rebase onto latest main and resolve conflicts. GitHub currently reports CONFLICTING/DIRTY, mainly because cli-manifest.json and adapter surface have moved substantially since this branch.

2. Drop the Z-Library files from this PR. This PR mixes two unrelated sites (weibo and zlibrary), and current main already has zlibrary/search, zlibrary/info, docs, manifest entries, and tests. Keeping the stale Z-Library implementation here would reintroduce old generic CliError / silent clamp patterns and duplicate already-landed work. If you still want a Z-Library follow-up, please open a separate PR based on latest main.

3. Rename weibo statuses before merge. statuses is Weibo/API jargon and does not fit the existing command family (user profile, post detail, feed personal timeline, search, comments). Please use weibo user-posts <id> as the clearer user task: list posts for a specific user. Update module path, command name, docs, manifest, function/test names, and PR body accordingly.

4. Add docs and manifest/index sync for the new Weibo command. Current PR only changes cli-manifest.json; docs/adapters/browser/weibo.md, docs/adapters/index.md, and README.zh-CN.md still do not mention the command.

5. Add contract tests for the Weibo command. At minimum cover: bad --limit / --page fail before fetch, screen-name-to-uid resolution, API error envelope -> typed failure, empty/malformed payload behavior, and row shape including id/mblogid/url round-trip into weibo post <mblogid>.

6. Replace silent argument coercion/clamping with upfront typed validation. Math.min(kwargs.limit || 15, 50) and kwargs.page || 1 silently accept 0, negative values, floats, and non-useful values. They should throw ArgumentError before browser/API work. This project has been removing silent clamps.

7. Tighten typed error boundaries. Non-2xx / API envelope errors should become CommandExecutionError; auth/login failures should become AuthRequiredError where distinguishable; true no-result should not silently return []. Avoid raw browser/evaluate errors leaking as untyped exceptions.

After the rebase + scope cleanup, I can re-review the Weibo-only diff in detail.