feat(schema): enforce no-whitespace in tags via pattern ^\\S+$

[mingcha-dev]

Summary

Adds hard schema-level enforcement that tags must not contain whitespace.

"items": {
-  "type": "string"
+  "type": "string",
+  "pattern": "^\\\\S+$"
}

Also updates description to state the rule + acronym-case convention.

Why

Tags containing spaces (e.g. "economic growth") break tokenizer/index semantics. Schema description has always said "mixed Chinese/English keywords" and examples always used hyphens (economic-growth), but there was no enforcement. This PR closes the gap.

Prereq

Merged PR #196 cleaned up 322 existing files (2967 space-tag violations → 0). Main is clean; this PR just locks the door behind us.

Validation

Ran make validate against current main — all files pass.

Three-way context

Ref: 2026-04-30 alignment (明鉴/墨子/明察). This is the 2nd of 2 back-to-back PRs.

1. PR chore: normalize whitespace in tags across all sources (keep Chinese) #196 cleanup (merged) ✅
2. This PR (enforcement)

Convention (documented in description)

• Must: no whitespace in tag strings
• Should (new entries): preserve case of acronyms (GDP/IPO/NNSA/CBIRC/etc.)
• Tolerated (historical): lowercase acronyms on main (from chore: normalize whitespace in tags across all sources (keep Chinese) #196 scope creep, accepted)

Checklist

• make validate passes on main
• Schema still allows mixed CN/EN
• No source file changes
• Description free of banned terms

[mingcha-dev]

protect-schema CI 规则：schema 改动必须直接 commit to main by maintainer，不能走 PR。

关闭此 PR。合适做法：

1. 在 main 分支直接 commit schema diff
2. 或老板配 bypass rule 给 trusted role

schema 改动内容（pattern + description 更新）已 ready，可由有权限人直接 commit。

[firstdata-dev]

Approved ✅

• Schema change 最小：items 加 "pattern": "^\\S+$" + description 标注

[firstdata-dev]

Approved ✅ schema 变更最小(items pattern + description)。本地 make validate 在 origin/main 全库通过。