Maina Lewis

rush is an AI pull-request reviewer that checks a change against what it was meant to do and against the team's own standards, instead of providing generic review. When a PR is opened, rush reads the ClickUp ticket linked in the description and the team's approved guideline pages. It then has a Meta model look into the diff through limited, read-only steps. It posts inline comments on the exact changed lines. Each comment shows the code evidence, the guideline or ticket requirement it breaks, and a suggested fix. Developers can reply in the thread to ask for a simpler explanation or to challenge a finding. rush re-checks a challenged finding and either keeps it with reasons or withdraws it. The problem: most AI reviewers don't know the ticket or the team's rules, so they flag style nits and miss "this doesn't do what the ticket asked" or "this breaks our tenant-isolation rule." In live tests, rush caught cross-tenant data access, a CSV formula-injection gap, floating-point money math, and an off-by-one against a ticket's acceptance criteria.