Context-Aware Retrieval
Concept
Documentation is for agents now. Structure knowledge so Claude (or any AI) can find what it needs.
Retrieval Paths
- Deployment setup →
03_Areas/Infrastructure/Deployment Pipeline - Database schema →
02_Projects/[Project]/Docs/Database Schema - Past decision → Meeting note where it was discussed
- Feature context → Project hub + related idea notes
Why Structure Matters
“Without structure you just have a pile of meeting notes. With structure you have a knowledge system that Claude can build on.”
The difference:
- Unstructured: AI searches randomly, misses context, hallucinates
- Structured: AI loads exactly what it needs, builds on prior work
Implementation Requirements
- CLAUDE.md / AGENT.md: Defines vault philosophy, folder structure, navigation patterns
- Folder READMEs: Deeper philosophy for each specific folder
- Consistent naming: Predictable paths for predictable retrieval
- Wikilinks: Explicit connections between related concepts
The Payoff
“Feature brainstorming becomes way better when Claude has all the context, all the ideas you’ve had before, all the decisions you’ve made, all the competitor research and all the user feedback.”