# ContextOps

> ContextOps is the practice of building, governing, evaluating, and repairing the context layer that AI agents use before they act.

It covers sources, policy, retrieval, packaging, evaluation, and repair.

Ambience's view: ContextOps is the production operating model for company context. Ambience grounds it in source-linked memory, permission-aware retrieval, MemoryOps checks, and audit evidence across approved agent runtimes.

## Related

- [ContextOps](https://ambience.sh/company-context/contextops)
- [MemoryOps](https://ambience.sh/company-context/memoryops)
- [Ambience vs internal RAG](https://ambience.sh/compare/internal-rag)