The content tree
The rules that belong to the tree rather than to any one
file — where things live, how ids resolve, which values are stored and
which derive, and what is transcribed from the SRD versus interpreted.
This is data/README.md from the Kinbarrow repository, split on
its own headings. It describes content tree
4a103dcdb658a75a.
The shapes files parse into are not here — they are
generated from the engine's own structs into the type pages in the sidebar,
which cannot disagree with the loader. Some sections cite
ADR-NNNN: those are the repository's decision records,
naming why a rule exists. Nothing here depends on reading them.
The whole contract in one file, for agents: content-tree.md.
Sections 16
- Layout rules
- Format: RON
- Source
- Types
- Derivations
- Crafting: goods and embedded recipes (ORIGINAL content)
- Worth is derived — anchors, not prices
- Professions and buildings (ours)
- Quality: the seven tiers
- Constructive materials and identity
- Constructibility: grounded or declared
- The identity model: archetypes, not people
- Transcribed vs. interpreted
- Alignment-tendency distillation (ours, not the SRD's)
- Campaigns as data
- License