brain-taxonomist

Determines filing paths for new brain pages by reading the active schema pack.

Updated Aug 17, 2026
One-click install
npx skills add https://github.com/AMC-JTC/gbrain-1 --skill brain-taxonomist-amc-jtc
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: brain-taxonomist
Source: https://github.com/AMC-JTC/gbrain-1/tree/main/plugin/skills/brain-taxonomist
Command: npx skills add https://github.com/AMC-JTC/gbrain-1 --skill brain-taxonomist-amc-jtc

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Deciding where a new knowledge-base page belongs is error-prone when directory conventions live in hardcoded tables or human memory. This Skill acts as a filing gate that reads the active GBrain schema pack via gbrain schema show --json and returns a reproducible, pack-driven path recommendation before any page is written. ## Core Features & Use Cases - Schema-driven filing decisions: Every recommendation comes from the active pack's page_types[].path_prefixes, never from a hardcoded directory table, so custom packs and per-source overrides are honored. - Drift detection: Runs periodic orphan scans with gbrain schema review-orphans and surfaces untyped pages when pack consistency degrades. - Ambiguity handling: Surfaces uncertain cases to the user instead of silently picking a closest-fit type, and signals EIIRP Phase 3 when no matching type exists. - Use Case: Before writing a meeting note about a new investor, ask where the page should be filed and receive a canonical path like people/<slug>.md with a one-line reasoning trail tied to the active pack version. ## Quick Start Ask the brain taxonomist where a new page about a person named Alice should be filed before writing it to the brain.

Frequently Asked Questions about brain-taxonomist

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I decide where to file a new brain page?

Consult the taxonomist before writing any new page. It identifies the primary subject type (person, organization, event, concept, media, or source), looks up the matching page_types entry in the active schema pack, and returns the canonical path prefix with a kebab-case slug.

Does the taxonomist use a hardcoded directory table?

No. Every decision is driven by gbrain schema show --json, so users on the gbrain-recommended pack or a custom pack authored via gbrain schema init get filing recommendations based on their own taxonomy, not baked-in defaults.

Can I use per-source overrides with multiple brains?

Yes. Pass --source <id> to every CLI call when working with a non-default source. Per-source overrides are tier 3 in the 7-tier resolution chain, and omitting the flag silently falls back to the brain-wide active pack.

What happens when no page type matches my content?

The skill does not pick the closest-fitting type. It signals EIIRP Phase 3 (SCHEMA CHECK) and suggests running gbrain schema detect followed by gbrain schema review-candidates to propose a new type through the schema-pack flow.

How do I detect taxonomy drift in my brain?

Run gbrain schema review-orphans --json to find pages with no type matching the active pack, and check gbrain doctor for the schema_pack_consistency check. When untyped pages exceed 10 percent, run the EIIRP Phase 3 flow to surface candidate types.

When should I skip the taxonomy check before writing?

Skip it when updating an existing page in place, appending to a Timeline section, or propagating meeting entities to existing pages. The check is mandatory only for new pages, bulk imports, and uncertain filing cases.