Resources
The vocabulary of AI asset management.
A growing library of concepts and practices behind PromptVerse. We publish the definitions first, and the deeper writing follows as we build.
Glossary
Core concepts.
- AI Asset
- A reusable, structured, testable and versioned AI instruction — a prompt, module, workflow or agent definition — managed as an object rather than as text.
- Module
- A reusable piece of content or logic that many assets link to, so it can be improved in one place.
- Wrapper
- Contextual instruction, metadata or processing behaviour applied around existing content without duplicating it.
- Variable
- A typed dynamic value inserted into an asset, letting one asset produce many personalized outputs.
- Evaluation
- A repeatable check of an asset's output against criteria you define, so changes can be measured instead of guessed at.
- Version
- A recorded state of an asset that can be compared, commented on, restored or published.
In progress
Topics we are writing about.
Prompt engineering as a discipline
Why production prompts need structure, tests and review — and what changes when they do.
planned
AI asset management
The practice of treating AI instructions as managed software assets with owners and history.
planned
Prompt optimization
How to define a target, generate candidates and validate improvement without fooling yourself.
planned
Model evaluation
Comparing models and providers against your own criteria rather than published benchmarks.
planned
Workflow design
Composing multi-step AI processes from reusable parts you can maintain centrally.
planned
AI agents
Managing agent instructions as versioned assets consumed through API, MCP or CLI.
planned
Stop managing AI in copy and paste.
Create a system for the prompts, workflows and agents your business depends on.