菜单
Use when designing knowledge bases, RAG, wiki, knowledge graphs, memory for agents, or "how should we store docs for humans vs agents". Produces a clear consumer/form/retrieval decision. Avoids buzzword soup. Do NOT use for ordinary app CRUD databases.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/AVIDS2/avid-skill --skill knowledge-architecture命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | knowledge-architecture |
| description | Use when designing knowledge bases, RAG, wiki, knowledge graphs, memory for agents, or "how should we store docs for humans vs agents". Produces a clear consumer/form/retrieval decision. Avoids buzzword soup. Do NOT use for ordinary app CRUD databases. |
Separate three decisions people often mash together: who consumes, what form, how retrieval works.
| Consumer | Optimize for |
|---|---|
| Humans | Readable structure, navigation, editability (Wiki / Markdown tree) |
| Agents | Chunkability, metadata, stable IDs, tool-friendly retrieval |
| Both (default for products) | Single source of truth humans edit + indexes agents query |
Default bias: coexistence — humans write Markdown/Wiki; agents get hybrid retrieval over the same source. Avoid a second shadow knowledge base unless required.
| Form | When |
|---|---|
| Markdown / Wiki | Most products; docs, runbooks, specs |
| Structured DB records | Strong schema entities (tickets, SKUs) |
| Knowledge graph | Relationships are first-class (org charts, law cites, multi-hop entity links) |
| Vector-only corpus | Large unstructured dump with weak structure needs |
MVP default: Markdown wiki (or /docs). Promote to graph when queries are relationship-heavy and hybrid RAG fails.
| Pattern | When |
|---|---|
| Full-text only | Small corpus, precise keywords |
| Vector RAG | Semantic questions over unstructured text |
| Hybrid (keyword + vector) | Default production RAG |
| Agentic RAG | Multi-step: plan → retrieve → critique → retrieve again |
| Graph RAG | Multi-hop entity/relation questions |
MVP default: hybrid retrieval. Use agentic RAG when single-shot retrieval quality is measurably insufficient.
## Consumers: human | agent | both
## Source form: markdown-wiki | db | graph | mixed
## Retrieval: fts | vector | hybrid | agentic | graph-rag
## Ingestion path
## Evaluation plan (5 real questions the system must answer)
## Non-goals
research-before-build, stack-defaultsreferences/knowledge-decision-tree.md