concept-family-explorer

Given a subject, concept-family-explorer finds the concepts in its conceptual family that are currently missing from an existing skill library or concept tree, fills the worthwhile ones to saturation by running /dr on each, and finishes by optimizing every skill it touched.

It is a gap-discovery orchestrator, not a researcher — per-concept research, skill authoring, and installation are /dr’s job. This skill decides which concepts are worth researching, in what order, and when to stop.

The five neighborhoods

Every subject gets decomposed into five neighborhoods before anything is scored:

Neighborhood Question it answers
Parent / super-domain What broader field is this a specialization of?
Siblings What sits at the same level under the same parent?
Children / sub-concepts What does this decompose into?
Adjacent / cross-over What neighboring domains overlap or interface here?
Frontier / emerging What is new, contested, or rising in this space?

Each candidate concept gets tagged HAVE, STALE, or GAP against existing coverage, then scored on five axes — Relevance, Usefulness, Novelty, Interest, Viability — before anything gets researched. Only gaps that clear the viability threshold get sent to /dr. The loop stops when two consecutive rounds of re-expansion turn up nothing new above threshold — evidence of saturation, not just a list running out.

Where it sits

Breadth, not depth: it maps everything around a subject and stops at one useful pass per neighbor. Its narrow inverse — going inside one concept instead of around it — is rabbithole. Once a family is mapped and worth compiling into a full llms-family reference set, full-suite is the next layer up.

Use it for: “what am I missing about X”, “map the conceptual family of X”, “what skills should I build around X”.

Not for: a concrete, already-named topic (/dr directly) · a cited research report with no skill-building (deep-research) · reorganizing an existing skill tree with no new research (skill-tree-architect).