thinkloom-openai-hackathon/.labyricorn/README.md

Labyricorn publishing content

This directory is the source of Thinkloom's project exhibition and development log on Labyricorn. The application repository remains the source of truth for this content; the Labyricorn site imports it as read-only content.

The content uses native Lektor records:

.labyricorn/
├── project/
│   ├── contents.lr
│   └── icon.png
└── devlog/
    ├── contents.lr
    └── <entry-slug>/
        └── contents.lr

Rules:

• project/contents.lr uses the project model.
• devlog/contents.lr uses the devlog model.
• Each published entry is a directory below devlog/ containing a contents.lr record using the devlog-entry model.
• Entry directory names are stable public slugs. Do not rename a published entry without arranging a redirect on the Labyricorn site.
• Dates use YYYY-MM-DD.
• source_commit is the full commit ID most directly associated with the entry. The site uses it to create a repository link.
• Images and other publishing attachments must stay inside .labyricorn/.
• Do not put credentials, generated builds, application state, or executable code in this directory.

The Labyricorn synchronization job reads content from an exact repository commit. Editing these records does not itself deploy the site; scheduled synchronization runs every 30 minutes and is the publication fallback. After an authorized push, an assistant with an approved refresh credential may dispatch the Labyricorn site's deploy.yml workflow once as documented in .labyricorn/AGENTS.md; the credential must never be stored in this repository.