coderabbit-review

What problem does it solve? CodeRabbit's GitHub App reviews pull requests automatically, but its findings arrive as check conclusions and review threads that must be waited on, triaged against a Scope Brief, and settled before merge — a process that is easy to get wrong by polling blindly, misreading pending states, or leaving unresolved threads behind a merge gate. ## Core Features & Use Cases - Status-protocol waiter script: Polls the CodeRabbit check with a wall-clock budget and emits a machine-readable STATUS line (TERMINAL_PASS, TERMINAL_FAIL, PENDING, NOT_CONFIGURED, ERROR) so callers branch on facts, not prose. - Scope Brief triage: Fetches all CodeRabbit review threads via paginated GraphQL and classifies each as blocking, immaterial, or deferred before handing dispositions to the rabbit-feedback-resolver skill. - Rate-limit degradation: Treats CodeRabbit quota limits as a waiver for future passes only, requiring every already-delivered thread to be settled before recording the gate as skipped. - Use Case: After pushing fixes to a PR, run the waiter in the background, triage the new threads it surfaces, dispatch the resolver with per-thread dispositions, and loop until the check passes with zero unresolved threads. ## Quick Start Wait for the CodeRabbit review on pull request 42, triage its findings against my Scope Brief, and resolve every thread until the check settles clean.