Capacitor/Review flows
The agent that wrote the code has the same blind spots when it reviews the code. A review flow hands your spec or diff to a separate reviewer — and loops until the findings are gone.
Without a second agent
You ask the agent that just wrote it to check its own work.
Clean handler — it follows the existing route style and reads well.
No issues. Good to merge.
One agent can’t catch the blind spot it just wrote in.
With a review flow
Your agent hands the diff to a hosted Codex reviewer and waits for the verdict.
The reviewer flagged the missing authorization check. I added an org-admin guard and re-submitted.
Second round came back clean — it’s reviewed and ready for the PR.
With a review flow, work gets checked, fixed, and re-checked before it’s done.
How it works
The kcap mcp flows server hands your spec or diff to a hosted Codex reviewer your daemon launches. It returns findings, your agent addresses them, and it re-reviews round after round until it signs off — the whole loop event-sourced on your server, so every round is on the dashboard. Read the mechanics in Review flows.
Start free, run a daemon on your repo, and let a separate agent check your agent's work before you merge.
Rather start a conversation? Talk to the team — we’re building with teams that already use coding agents.
Built by the team behind KurrentDB — event streams in production are what we do. Coding agents just produce a new kind.