Answers instead of archaeology
Who publishes OrderPlaced? What breaks if we change it? Your team searches the catalog instead of grepping repos and asking around.
Turn your repo into documentation.
One prompt. Your agent reads the code and writes the catalog.
Nothing to install.
Read the EventCatalog code-to-catalog skill at
https://github.com/event-catalog/skills/tree/main/skills/code-to-catalog
(fetch SKILL.md and the files in references/), then follow it on this repository.Prefer it installed?
What your team gets
The prompt takes minutes. What you're left with is an architecture your whole team can search, trust, and keep current.
New engineers ship sooner
The architecture is written down, with owners attached. Onboarding stops depending on who happens to be free that week.
Changes stop landing as surprises
Producers and consumers are mapped, so you can see who a schema change affects before you make it, not after it reaches production.
A real map of your systems
Visualise how domains, services, messages, and channels connect — the diagram nobody has kept current since the last redesign.
Docs that live beside the code
Markdown in your repo, reviewed in pull requests. It stays current because it changes the same way your code does.
Context your AI agents can use
The same catalog your team reads becomes grounding for coding agents, so they reason about your architecture instead of guessing.