Taxonomy and Retrieval

A library without a taxonomy is a pile. Every prompt is in there somewhere, but finding the right one requires either knowing exactly what it is called or searching through everything. The taxonomy is the index that makes the library useful rather than merely comprehensive. The right taxonomy for a prompt library organizes prompts the way the people who use them think about them — by function, by domain, by agent, or by use case.

1. Dimension 1: Function What does the prompt do? Generate, extract, classify, summarize, compare, translate, format, evaluate. Function-based organization answers "I need a prompt that does X."
2. Dimension 2: Domain What business domain does the prompt operate in? Sales, marketing, legal, finance, operations, customer success. Domain-based organization answers "I need a prompt for use in X department."
3. Dimension 3: Use Case What specific workflow does the prompt support? Follow-up email, risk assessment, competitive brief, invoice extraction, onboarding checklist. Use case organization answers "I need a prompt for X task."
4. Dimension 4: Agent or Author Which agent or team member owns and maintains this prompt? Ownership metadata answers "Who do I contact when this prompt needs to change?"

Multi-dimensional tagging is more useful than hierarchical organization for prompt libraries because most prompts belong to multiple categories. A sales follow-up email prompt is in the "generate" function category, the "sales" domain, and the "email" and "follow-up" use cases simultaneously. A flat taxonomy with multi-dimensional tags is searchable from any angle. A rigid hierarchy requires knowing which branch the prompt was filed under.