Researcher workbench

Contribute a node

Draft a new atom of knowledge against the live corpus. The checks on the right mirror the kernel's structural rules as you type — but admission is the kernel's alone: place the JSON in content/nodes/, run python -m ouro verify, then python -m ouro.export_web, and open a pull request. Nothing enters the graph without passing verification.

Kind

Name (stable identity slug)

References use this name forever — choose carefully.

Title

Statement

Domain

Foundation

Dependencies — what this proof uses (0)

Proof

References (one per line, optional)

Structural checks (mirror of the kernel)

✗Name is a valid slug — lowercase letters, digits, hyphens — e.g. lem-my-result

✗Name is unique in the corpus — no collision

✓Conventional prefix for kind

✗Statement is non-empty — the statement is the atom — it cannot be blank

✗Title is non-empty

✗Proven kind carries a proof — theorems/lemmas/propositions/corollaries must include a proof

✗Proven kind depends on something — a proof that uses nothing proves nothing new

✓No self-dependency

✓All dependencies exist

✓Foundation is strong enough for dependencies

✓Cannot create a cycle

Resolve the ✗ items — the kernel would reject this draft.

Node JSON

{
  "kind": "lemma",
  "name": "",
  "title": "",
  "statement": "",
  "domain": "number-theory",
  "foundation": "peano",
  "dependencies": [],
  "relations": [],
  "proof": null,
  "references": []
}

Then: content/nodes/your-node.json → python -m ouro verify → python -m ouro.export_web → PR. The content hash is computed by the store on save.