Translate AI feature docs and assistant API links to es, fr, zh

[mintlify]

Summary

Translates changes from commits since the last translation PR (#4647) into Spanish, French, and Chinese:

• ai/llmstxt.mdx: Added .well-known URL mention, new HTTP headers discovery section, .md extension note in page links, restructured code example (split OpenAPI/AsyncAPI specs)
• ai/markdown-export.mdx: Added new "Accept header" section for programmatic Markdown fetching
• ai/skillmd.mdx: Removed "(with caching disabled)" from reverse proxy note, added new "Skills discovery endpoints" section with index and individual file endpoints
• api/assistant/create-assistant-message-v2.mdx: Fixed internal link from /api-reference/ to /api/
• api/assistant/create-assistant-message.mdx: Fixed internal link from /api-reference/ to /api/
• Also replaced HTML entities (') with real quotation marks in FR and ES assistant files

---

Note

Low Risk
Low risk documentation-only changes (translations/wording updates) with no runtime code changes; primary risk is broken internal links or inaccurate guidance.

Overview
Updates the Spanish/French/Chinese AI docs to reflect recent llms.txt/llms-full.txt behavior: adds .well-known locations, documents HTTP discovery headers, notes .md page-link extensions, and refactors the example to split OpenAPI vs AsyncAPI specs.

Adds a new section describing programmatic Markdown retrieval via the Accept: text/markdown (or text/plain) header, and expands skill.md docs with /.well-known/skills/ discovery endpoints (index + per-skill file) while simplifying reverse-proxy guidance.

Fixes internal assistant API doc cross-links from /api-reference/... to /api/... and cleans up quote/entity formatting in the localized assistant pages.

^{Written by Cursor Bugbot for commit 7afe01f. This will update automatically on new commits. Configure here.}