LinkCatalog

LinkCatalog is the durable workspace at /home/koni/workflows/linkcatalog. It is the one part of the lab that is not optional: every other system either reads from it, writes to it, or publishes something derived from it.

The directory is ordinary. It is a set of files on disk, organised by hand and by cron job, that together form a personal record of every link worth keeping — bookmarks, Discord drops, GitHub stars, long reads, public notes, and the small trails that connect them. The structure is meant to be human-readable first and machine-parsed second, not the other way around.

What it holds

• Links, each with a source, a date, a set of tags, and a short note on why it was kept.
• Trails, which are ordered links plus prose: the path of an idea through several sources.
• Indexes, generated from the links: by tag, by source, by date.
• Pending drops, links acknowledged but not yet catalogued — the ingest queue.

What it is not

LinkCatalog is not a browser history and not a read-later queue. It is a selection. Things enter it because they were worth keeping; they stay because they are still worth finding.

Connections

• Discord link ingest acknowledges dropped links into the pending queue.
• GitHub stars sync periodically imports starred repositories as links.
• Citta Publishing System reads from this path and publishes the catalogue to the web.
• Research Trails are stored here as first-class artefacts.