You say. Your agent knows.
Three everyday asks. Ask your agent to use CKB, or add it to your agent's instructions, so it gathers evidence instead of guessing.
Refactor an auth flow
“Refactor the login flow to use the new session token format.”
Before touching code, your agent asks CKB for every caller, test, and contract that touches the auth flow. It refactors with that list in hand instead of guessing at what depends on it.
Why does changing this type break the API?
“Why does changing this type break the API?”
Your agent traces the type through every function signature and consumer that depends on it, and gets back the exact points where the change breaks something — not just a guess at "probably somewhere in the API layer."
Review a PR
“Review this PR.”
Instead of re-reading every changed file from scratch, your agent gets structural findings back — breaking changes, dead code, test gaps, secrets — and writes the review from that.
More prompts
The Prompt Cookbook has more examples of everyday asks and what your agent gets back from CKB for each one.