A role inside a multi-agent review swarm In this channel's material, a correctness reviewer appears as one of two reviewer roles in a custom multi-agent review swarm used to vet generated code before the human engineer looks at it [^7d2a9b32-b281-5828-9eb1-bf812129ae58@426]. How it fits into the workflow The swarm review skill works in steps: - Divide the change set into domains — for example frontend and backend. - Delegate per domain to two reviewer roles: a correctness reviewer, hunting for errors, and a code maintenance reviewer, looking for patterns — code to reuse, existing patterns that should be followed, and things that will fall apart at scale. - Run every pass on both GPT and Opus — two models per check. - A lead agent collects all findings and turns them into a report of what needs to be done. - Iterate until it's good — the agent repeats the loop until the findings are resolved [^7d2a9b32-b281-5828-9eb1-bf812129ae58@426]. Why the split matters Big change-sets need more than bug-hunting — they also need checks for maintainability. The correctness reviewer and the code maintenance reviewer are deliberately complementary: one asks does this code do the right thing, the other asks will this code hold up [^7d2a9b32-b281-5828-9eb1-bf812129ae58@426]. - A correctness reviewer is a reviewer-agent role focused on finding errors in code. - It is one of two roles (alongside the code maintenance reviewer) in the swarm review skill, assigned per domain of a change set. - Findings from every pass feed a lead agent's report, and the loop repeats until findings are resolved.