Reusable artificial intelligence context packages information that agents repeatedly need, such as repository conventions, architecture, domain rules, examples, tool guidance and prior decisions. It can be stored in documentation, skills, indexed knowledge or other versioned resources.
Shared context reduces repeated prompting and human correction, but it needs ownership and maintenance. Outdated or conflicting material can spread the same mistake across many tasks, so teams should test changes and make provenance visible.
Acronyms and aliases
shared agent context synonymreusable AI context variant
Related terms
Frequently asked questions
What belongs in reusable AI context?
Stable conventions, architecture, domain knowledge, examples, constraints and tool instructions that apply across multiple tasks belong there.
How should reusable AI context be maintained?
Teams should version it, assign ownership, test important changes and remove contradictions through a reviewed update process.
