AgentSkillsCN

wiki-ontology-design

当用户希望以分层本体论为架构，从模糊的主题云逐步细化为清晰的地图，从而设计维基结构时使用。当您需要为全新语料库规划信息架构时，此技能同样适用。若需对维基页面进行多源研究，请参考 multi-source-research。

SKILL.md
--- frontmatter
name: wiki-ontology-design
version: 1.0.0
description: "When the user wants to design a wiki structure by building and refining a layered ontology from a fuzzy topic cloud to a sharp map. Also use when planning information architecture for a new corpus. For multi-source research into wiki pages, see multi-source-research."

Wiki Ontology Design

Use this skill to design the wiki structure for a new corpus.

Workflow

  1. •Fuzzy scan: extract top terms and co-occurrences.
  2. •Broad domains: 3-7 buckets that define the worldview.
  3. •Subtopics: 6-15 per domain, remove overlaps.
  4. •Canonical map: classify into frameworks, cases, people, episodes.
  5. •Approve: write approved.json and generate scaffolds.

Script

bash
python3 scripts/ontology_refiner.py --input phase1_domains.json --output approved.json

References

  • •references/blur-to-focus.md
  • •references/ontology-schema.json

Related Skills

  • •Upstream: multi-source-research
  • •Enhanced by: semantic-chunking-rules
  • •Feeds into: podcast-wiki-pipeline