infinite-brain-os: Command for Claude Code

/build-knowledge-base

Launch a real namespace build or migration in plain English.

This command is for cases where an operator wants to say something like:

Build a new namespace called personal-health from <legacy-root>\legacy-notes\Health, do not migrate the legacy database, and refactor the rest into Infinite Brain style.

The command does not require rigid mode= or exclude= flags. The operator describes the source, the target namespace, and any constraints in normal language. The command then routes that request through the V2 namespace machinery.

When to use this

• Building a brand-new real knowledge namespace under knowledge/
• Migrating an existing corpus, folder, or PKM subtree into Infinite Brain V2
• Testing whether the current skills, agents, workflows, validator, and doctrine are good enough to convert a real source corpus
• Refactoring semi-structured notes into canon, synthesis, support, and profile-aware structure

When NOT to use this

• You only need a lightweight namespace registry stub. Use /create-namespace.
• The inbound material is still raw intake and should go to root intake/ first.
• The work is mostly runtime state, secrets, operational queue data, or databases that do not belong in a knowledge namespace.
• The request is to add or revise one node inside an existing namespace rather than create or migrate the namespace itself.

Operator input style

The operator may describe the job in natural language. The command should extract, from the request, as much of the following as is available:

• source location or source set
• target namespace slug
• target repo
• what to exclude
• what canon expectations matter
• what the namespace should be used for
• any known constraints, sensitivities, or live-truth risks

Do not force the operator into a fixed flag schema when plain-English instructions are sufficient.

What this command does

1. Reads the operator request and determines whether this is:
   • a born-V2 namespace build, or
   • a migration of an existing knowledge corpus into a new namespace.
2. Reads the V2 doctrine and operative contract:
   • AGENTS.md
   • CLAUDE.md
   • _system/README.md
   • _system/namespace-profiles.md
   • knowledge/ai-architecture/INDEX.md
   • knowledge/ai-architecture/canon/core-doctrine.md
   • knowledge/ai-architecture/canon/problem-to-architecture.md
3. Chooses the namespace profile from first principles using [[skill-build-namespace]].
4. Creates or updates the registry entry under _system/namespaces/.
5. Scaffolds the namespace with the shared base and profile-additive folders.
6. Routes the actual build through [[skill-build-knowledge-base]] and [[workflow-build-knowledge-base]], which in turn sequence the lower-level V2 machinery:
   • [[skill-build-namespace]] for new namespace scaffolding
   • [[skill-migrate-legacy-knowledge-to-v2]] or [[workflow-legacy-knowledge-migration]] for corpus refactoring
   • [[skill-canonize-namespace]] for compressed canon
   • [[skill-refine-namespace-index]] for the retrieval router
7. Atomizes broad source material into digestible nodes before building higher-order summaries, so future agents can ingest the graph selectively rather than only reading a few large notes.
8. Preserves provenance in support/, puts derived interpretation in synthesis/, and keeps canon compressed and agent-usable as the main-ideas layer future agents should ingest first.
9. Reviews related namespaces for additive node, synthesis, canon, or router updates triggered by the new build.
10. Runs bash _system/validate.sh plus [[workflow-namespace-lint-review]] before closing.
11. Produces a short build or migration report describing:

• chosen profile
• sources used
• major atomization decisions
• canon surfaces created
• synthesis surfaces created
• cross-namespace changes made or deferred
• support receipts created
• validator/lint results
• any open questions or human approval gates