LLM-Wiki Model Policy

Goal

Create a practical model/data policy for ingest, query, lint, capture and publication workflows.

When to use

• The user asks which models or providers are allowed to process a given wiki source or folder.
• The user wants to decide what must stay local-only versus what can go to a cloud model.
• The user is setting up or revising the cheap/local vs. heavy/cloud model split across ingest, triage, synthesis, query, lint and embedding tasks.
• The user asks how to record model provenance (which model produced a page) in wiki frontmatter.
• The user needs escalation rules for low-confidence outputs or sensitive raw sources before publication.

Inputs