concept-synthesis

Deduplicate concept stubs, assign tiers, and generate a tiered intellectual map.

1|Updated May 9, 2026
One-click install
npx skills add https://github.com/weiping/gbrain-cn --skill concept-synthesis
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: concept-synthesis
Source: https://github.com/weiping/gbrain-cn/tree/main/skills/concept-synthesis
Command: npx skills add https://github.com/weiping/gbrain-cn --skill concept-synthesis

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Raw concept stubs created during ingestion quickly become duplicated, flat, and unconnected, leaving you without a coherent intellectual map or traceable idea evolution.

Core Features & Use Cases

  • Deduplicate and merge concepts: Combine near-duplicates deterministically (string/title overlap and semantic duplicate detection) into canonical concept pages while preserving aliases and timelines.
  • Tier concepts into a hierarchy: Score each canonical concept by frequency, timespan, breadth, and available engagement signals, then label it as T1 Canon, T2 Developing, T3 Speculative, or T4 Riff.
  • Synthesize and connect evolution: Generate verifiable, quote-grounded synthesis for T1/T2 concepts and cluster tiered concepts into named intellectual domains, producing a master concepts/README map.

Quick Start

Tell your agent to run concept synthesis to dedupe concept stubs, assign tiers, and produce the concepts intellectual map.

Frequently Asked Questions about concept-synthesis

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

FAQPage Schema
How do I deduplicate concept stubs generated during knowledge base ingestion?

To deduplicate concept stubs, you can run concept synthesis to deterministically merge near-duplicates using string and title overlap alongside semantic duplicate detection, while preserving aliases and timelines.

What is the best way to organize a flat knowledge base into a tiered intellectual map?

Organizing a flat knowledge base into a tiered intellectual map requires scoring canonical concepts by frequency, timespan, and breadth, then assigning them to T1 Canon, T2 Developing, T3 Speculative, or T4 Riff tiers.

How do I trace idea evolution across multiple sources over time?

Tracing idea evolution across sources over time is achieved by generating verifiable, quote-grounded LLM synthesis for T1 and T2 concepts and clustering tiered concepts into named intellectual domains.

When do I need canonicalization and tier scoring for my knowledge graph?

You need canonicalization and tier scoring when your ingestion pipelines generate many near-duplicate concept pages, leaving your knowledge base flat and unconnected without a coherent intellectual map.