Skip to main content

Output Directory Layout

After running /wiki-compile, your configured output directory contains:
All files are plain markdown with Obsidian-style [[wikilinks]]. Open wiki/INDEX.md in Obsidian and you’ll see the full knowledge base with bidirectional links to source files. Set "link_style": "markdown" in .wiki-compiler.json if you prefer standard links.

INDEX.md — The Entry Point

INDEX.md is the file your agent reads first at session startup. It lists every topic and concept with source counts, last-updated dates, and status, so Claude knows at a glance what the wiki covers and which articles to read for the current task.

Topic Articles

Each topic article covers everything known about one subject, synthesized from all contributing source files. Topic files live in topics/ and use lowercase-kebab-case filenames (e.g., topics/d1-retention.md). The frontmatter records compilation metadata:
Every section in the article carries a coverage indicator so you and your agent know how well-sourced each part is:
The Sources section at the bottom of every article lists backlinks to every raw file that contributed — so you can always trace synthesized claims back to their origins.
The default article structure (Summary, Timeline, Current State, Key Decisions, Experiments & Results, Gotchas, Open Questions, Sources) is a starting point. During /wiki-init, the compiler proposes a structure tailored to your domain. You can edit article_sections in .wiki-compiler.json at any time.

Concept Articles

Concept articles live in concepts/ and capture patterns that span three or more topics. They are interpretive — they answer “what does this pattern mean across the project?” rather than just recording what happened in each topic. The compiler discovers concepts automatically after each full compile. You can also seed them in schema.md if you know what patterns you want tracked.

schema.md — The Wiki’s Constitution

schema.md is generated on first compile and defines the rules the compiler follows on every subsequent run. It covers:
  • Topics — every topic slug and a one-line description of what it covers
  • Concepts — cross-cutting patterns and the topics they connect
  • Article structure — section names, order, and coverage tag format
  • Naming conventions — slug format (lowercase-kebab-case), date format (YYYY-MM-DD), link style
  • Cross-reference rules — topics that share 3+ sources should reference each other
  • Evolution log — a chronological record of every schema change
You can edit schema.md directly. Rename a topic slug, merge two topics, or add a convention — the compiler reads it before each run and applies your changes. New topics discovered during compilation are added automatically with a log entry.
Don’t delete topics from schema.md without also removing the corresponding file from topics/. Run /wiki-lint after schema edits to catch any drift between the schema and the actual wiki files.