Main
- Start with a minimal tool surface and expand only after you have observability and reset/recovery procedures.
- Store secrets in the mechanism the installer sets up; avoid baking keys into compose files or images.
- Pin versions for reproducible environments once you move from demo to team usage.
README excerpt (source)

# E.D.D.I — Multi-Agent Orchestration Middleware for Conversational AI
[](https://www.bestpractices.dev/projects/12355) [](https://securityscorecards.dev/viewer/?uri=github.com/labsai/EDDI) [](https://app.codacy.com/organizations/gh/labsai/dashboard?utm_source=github.com&utm_medium=referral&utm_content=labsai/EDDI&utm_campaign=Badge_Grade)
[](https://github.com/labsai/EDDI/actions/workflows/ci.yml) [](https://github.com/labsai/EDDI/actions/workflows/codeql.yml)  FAQ
Q: Does it require Docker? A: The README quick start requires Docker and uses Docker Compose via the installer.
Q: How do updates work?
A: The installer creates an eddi CLI wrapper; eddi update pulls and restarts containers.
Q: Safe first use case? A: Start with non-sensitive test data and a read-only tool setup before connecting real systems.