Index and log conventions

Paths and filenames come from the wiki schema. This document describes behaviors that work across layouts.

Index (content catalog)

Purpose: Let the agent (and human) find pages without embedding search. At small-to-medium scale, reading the index first is often enough.

Typical contents:

• Grouping by category (entities, topics, sources, overviews)—exact names live in schema.
• Each entry: wiki link, one-line summary, optional metadata (date, source count) if the schema uses frontmatter or tables.

Update rule: After every ingest that adds or materially changes wiki pages, update the index so it stays a complete catalog of wiki pages the schema cares about.

Log (append-only timeline)

Purpose: Chronological audit trail: what happened, when, and which files moved.

Discipline:

• Append only. Do not rewrite old entries to “clean up”; add a corrective entry if needed.
• Stable heading pattern so Unix tools and scripts can parse lines:

## [YYYY-MM-DD] ingest | Short title
## [YYYY-MM-DD] query | Short label
## [YYYY-MM-DD] lint | Short label

Body: Bullets listing touched wiki paths, new sources referenced, or lint findings summary.

Grep examples (adjust path to schema):

grep '^## \[' wiki/log.md | tail -10
grep 'ingest' wiki/log.md

Optional: YAML frontmatter on wiki pages

If the schema allows, use frontmatter for date, tags, sources—enables Dataview-style queries in Obsidian and similar tools. Do not require it unless the schema says so.