Every canonical document is classified along 5 dimensions: IP layer, system layer, methodology role, audience, and lifecycle stage. This is the taxonomy that turns a pile of docs into a coherent knowledge architecture.
Together these dimensions answer five questions about every doc: How protected is it? · Where does it operate in the stack? · What is its function? · Who is allowed to see it? · Where is it in its lifecycle?
The most consequential matrix. Tells you which docs need the strongest IP protection at which layer of the stack.
Tells you what kind of function each doc serves and who can see it. Catches misclassifications: e.g., a narrative doc tagged "internal" is likely an underused asset.
The classification system above sits on top of the IP architecture defined in canon doc BA-IP-Compartmentalization-Architecture-V2.0.html. Under V2 canon (2026-06-12), the relationship is:
Single source of all mathematical computation. LESRI engine, scoring, classification, phase state machine, safety gates. Trade-secret algorithms live here only. Classified system layer = computation. IP layer L4 (the most protected).
Master data store. Stores inputs, outputs, configuration values — never computes. The Worker writes through directly to Supabase. Classified system layer = storage. IP layers: L3 (schema) / L4 (algorithm-derived rows).
One-way downstream mirror for operational-workflow UIs only. Zero AI agents, zero algorithm logic. Coach-facing views and assessment intake. Classified system layer = presentation. IP layers: L1-L3. Selective Make.com flows (cost-analysis gated) shim where needed.
Decision tree for when a new canon doc is added.
Every canonical document, classified. Click a row to open the underlying file.