Agent files
Plain-text files
Anything that can read text can read this site. These files carry the same published content as the pages and the MCP server, with no markup around it.
The files
| Path | What it is | Type |
|---|---|---|
/llms.txt | A map, not the content: status and contact, then every case study and craft piece with a link and its Markdown address. Start here. | text/plain |
/llms-full.txt | The complete content in one request: every case study, the about page and professional experience. | text/plain |
/about.md, /experience.md, /craft.md | The Markdown version of each of those pages. | text/markdown |
/case-studies/{slug}.md | One case study. The same bytes the MCP server returns for get_case_study with detail set to full. | text/markdown |
/practice/{slug}.md | One practice note, with the evidence it declares. | text/markdown |
/craft/{slug}.md | One craft piece. | text/markdown |
Slugs are listed in llms.txt and returned by list_case_studies. The pages themselves link to their Markdown version too, so an agent that starts on a case study can find its twin.
For questions rather than whole documents, the MCP server searches and filters the same content.
Fetch one
The index first:
curl -s https://jason.sonderman.info/llms.txtThen one case study as Markdown:
curl -s https://jason.sonderman.info/case-studies/arcos-harmony-design-system.md