Improve AI docs discoverability + canonicalize AI artifact origin

[rossgalloway]

Summary
This PR improves how users and agents discover Yearn’s AI-readable docs resources and ensures generated artifact metadata is canonical by default.

What changed

• Added a new docs page at /getting-started/llm-friendly-docs:
  • Canonical AI endpoints (/llms.txt, /ai/manifest.json, /ai/docs.jsonl, /ai/raw/)
  • Example prompt for LLM use
  • Install instructions for both Codex and Claude Code for the yearn-docs-site-query skill
• Added LLM Friendly Documentation links in all sidebars:
  • User docs: directly under FAQ
  • Dev docs: above Deprecated Products
  • DAO docs: bottom of sidebar
• Updated scripts/generateAiArtifacts.ts:
  • siteOrigin now defaults to https://docs.yearn.fi
  • no longer auto-uses VERCEL_URL
  • still supports explicit override via DOCS_URL
• Updated README AI exports section to document DOCS_URL override behavior.
• Includes a timestamp refresh in scripts/fetchedAddressData.json (timeLastChecked).

Why

• Prevents /llms.txt and AI manifest from unintentionally pointing to preview hosts.
• Makes the AI/LLM entry points easy to find for users across docs sections.
• Gives copy/paste install paths for the hosted docs skill in both Codex and Claude Code.

Validation

• bun run build passes successfully.
• New route is generated: build/getting-started/llm-friendly-docs/index.html.
• Sidebar placements verified in all three sidebar configs.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
yearn-docs	Ready	Preview, Comment	Feb 24, 2026 9:51pm

[engn33r]

lgtm