brain-taxonomist

name	brain-taxonomist
description	Decide where new local brain pages belong before they are written, using repo conventions and any available schema guidance.

Brain Taxonomist

Use this skill before creating a new brain/ page when the right category, slug, or page type is uncertain.

Upstream GBrain reads the active schema pack. This Codex port first uses the repo-local brain layout and can consult gbrain schema show --json when the upstream CLI is available.

Workflow

1. Identify the primary subject:
   • person
   • company or organization
   • concept
   • source material
   • idea
   • meeting
   • report
   • task or workflow
2. Prefer existing local conventions:
   • brain/people/ for people
   • brain/concepts/ for reusable concepts, skills, tools, and projects
   • brain/sources/ for source documents and imported references
   • brain/ideas/ for original proposals or sketches
   • brain/.raw/ for preserved raw input
3. If gbrain is installed, inspect active schema metadata before inventing a new location.
4. Produce a filing recommendation:
   • path
   • slug
   • page type
   • required backlinks or citations
5. If no existing category fits, recommend schema-author before writing.

Use routing-eval.jsonl as the deterministic routing fixture. Each JSON object must identify an expected packaged skill; any ambiguous_with targets must also exist in the package.

Guardrails

• Do not hardcode a new directory if an active schema says otherwise.
• Do not create the page unless the calling task asked you to write it.
• Do not refile existing user pages without explicit permission.