How eriic-builds/fieldnotes works
Fieldnotes is a documentation-authoring framework for producing source-backed Markdown collections that AI agents can discover and use. The main workflow is to research a subject, organize and author focused pages, then expose the collection through the generated llms.txt index and a searchable website. The README also describes validation and publishing, but those are maintenance and deployment support rather than core authoring stages. Wiring beyond the documented workflow is not inferred from the sampled sources.
Main entry points: docs Markdown corpus · build_llms.py Index generator · topics Topic collection · mkdocs.yml Searchable site
Components 11
Authoring guidance
- Agent instructions
AGENTS.md - Research workflow
docs/guides/research-workflow.md - Authoring guide
docs/guides/authoring.md - Maintenance guide
docs/guides/maintenance.md
Documentation content
- Starter pages
docs/starter-pages - Topic collection
docs/topics - Markdown corpus
docs
Discovery and browsing
- Index generator
scripts/build_llms.py - Agent index
llms.txt - Searchable site
mkdocs.yml
Other components
- Authoring agent
Connections 9
- Authoring agent → Agent instructions: follows
- Authoring agent → Research workflow: researches with
- Authoring agent → Authoring guide: authors with
- Authoring agent → Starter pages: adapts
- Authoring agent → Topic collection: creates
- Topic collection → Markdown corpus: stores pages
- Index generator → Markdown corpus: reads pages
- Index generator → Agent index: generates
- Searchable site → Markdown corpus: renders