Domain Library
Interchangeable knowledge domains
The challenge system supports multiple knowledge domains through an interchangeable library. Each domain is a self contained bundle that defines everything needed to generate challenges in that domain.
How domains work
- The coordinator loads approved domain bundles at startup.
- Each challenge is generated from a spec inside a bundle, seeded by the epoch seed.
- Bundles ship their own validators, so scoring logic travels with the domain.
- Adding a domain does not require changing the mining protocol.
Bundle contents
| Part | Purpose |
|---|---|
| Generator | Produces challenge instances from a seed and difficulty band |
| Constraint set | Machine checkable rules the answer must satisfy |
| Validator | Deterministic replay function returning pass or fail with reason |
| Corpus reference | Any source material, content addressed so it cannot drift |
| Manifest | Domain id, version, class coverage, license |
Launch domains
The protocol opens with general reasoning domains that need no proprietary corpus: logical inference chains, document synthesis, and structured extraction. Additional domains are added by proposal.
Determinism requirement
A domain is only approved if its validator is deterministic: the same artifact and the same seed must always produce the same verdict on any machine. Non deterministic scoring cannot be independently replayed, and unreplayable scoring is not verifiable.