Handbook
Concept documentation
Operators and authors who need to understand what an intelligence means and how it runs without reading pack source trees.
Updated
Who is this for?
Operators and authors who need to understand what an intelligence means and how it runs without reading pack source trees.
What problem does it solve?
Domain packs contain many named intelligences (e.g. reasoning, planning, emotional regulation in Human-Life). The concept dual-wiki pairs two views per intelligence identity so definitions and execution stay aligned.
What can you do?
| Surface | You get… |
|---|---|
| Semantic | What the intelligence means — scope, capabilities, invariants |
| Execution | How it runs — tasks, validators, substrate reuse |
Browse the semantic index for definitions and the execution index for runtime overlays. Human-Life v4.1 ships 39 paired overlays; MathGenesis adds spine concepts.
What can you not do?
- Edit semantic or execution pages independently without parity checks — identity fields must match across both surfaces.
- Treat concept pages as operator marketing copy — they are contract-deep; this page orients you to the pattern.
Limitations
Not every cataloged intelligence has full concept wiki depth yet. Math spine overlays are newer scaffolds. Parity and maturity are enforced in maintainer tooling — see the system page.
Next steps
- Packs and domains
- System depth: paired system page
- Concept wiki hub