feat(seo): add llms-full.txt with the full text of every document

[ahnafnafee]

Summary

llms.txt is only an index, so an assistant that wants the actual content has to crawl each URL or hit the per-post markdown API. This adds a companion llms-full.txt following the llmstxt.org convention: every blog post, research entry, and portfolio project inlined verbatim under an H1 and a URL: line.

• 26 documents, ~171 KB, generated by the same scripts/generate-llms-txt.js that already writes llms.txt.
• MDX bodies carry no H1 of their own (the page template renders the frontmatter title), so one H1 per document keeps the concatenated heading hierarchy intact.
• llms.txt now points at it from both ## Discovery and # How to Use This Site.
• Both files are tracked and regenerated by the build, so the committed copies are the fallback and the build output is what ships.

CI

Adds a git diff --exit-code check on the pair after the build, so a content edit committed without a regenerate fails the run. This is the same class of silent drift that left llms.txt six posts stale.

Caveat on impact

llms-full.txt is an unratified community convention. No major provider has confirmed using it as a retrieval or ranking input, and Google has said publicly it does not use llms.txt. This is cheap insurance if adoption happens, not a measurable ranking change today.

Test plan

• yarn generate:llms writes both files; 26 documents, 171 KB
• yarn build regenerates both as part of the chained build
• git diff --exit-code on the pair is clean after a build (the CI gate passes)
• yarn lint, yarn validate:sitemap, yarn validate:json-ld
• CI green
• After deploy, /llms-full.txt returns 200 text/plain

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
ahnafnafee-dev	Ready	Preview, Comment	Jul 28, 2026 3:41pm