Skip to main content Apply production-ready CodeRabbit SDK patterns for TypeScript and Python.
Use when implementing CodeRabbit integrations, refactoring SDK usage,
or establishing team coding standards for CodeRabbit.
Trigger with phrases like "coderabbit SDK patterns", "coderabbit best practices",
"coderabbit code patterns", "idiomatic coderabbit".
npx skills add jeremylongshore/claude-code-plugins-plus-skills --skill coderabbit-sdk-patterns ai automation claude-code devops mcp ai-agents
CodeRabbit SDK Patterns
Overview
CodeRabbit does not have a traditional SDK. You interact with it through .coderabbit.yaml configuration, PR comment commands (@coderabbitai), and the GitHub/GitLab API to process its review output. These patterns show how to automate around CodeRabbit reviews programmatically.
Prerequisites
CodeRabbit installed on repository (see coderabbit-install-auth)
GitHub CLI (gh) or GitHub API access via personal access token
Node.js 18+ for automation scripts
Instructions
Step 1: Fetch CodeRabbit Reviews via GitHub API
// scripts/fetch-coderabbit-reviews.ts
import { Octokit } from "@octokit/rest";
const octokit = new Octokit({ auth: process.env.GITHUB_TOKEN });
async function getCodeRabbitReview(owner: string, repo: string, prNumber: number) {
const reviews = await octokit.pulls.listReviews({ owner, repo, pull_number: prNumber });
const coderabbitReview = reviews.data.find(
(r) => r.user?.login === "coderabbitai[bot]"
);
if (!coderabbitReview) {
console.log("No CodeRabbit review found yet. Review typically takes 2-5 minutes.");
return null;
}
return {
state: coderabbitReview.state, // "APPROVED" | "CHANGES_REQUESTED" | "COMMENTED"
body: coderabbitReview.body, // Walkthrough summary
submittedAt: coderabbitReview.submitted_at,
};
}
Step 2: Extract Line-Level Comments async function getCodeRabbitComments(owner: string, repo: string, prNumber: number) {
const comments = await octokit.pulls.listReviewComments({
owner, repo, pull_number: prNumber,
});
const coderabbitComments = comments.data
.filter((c) => c.user?.login === "coderabbitai[bot]")
.map((c) => ({
file: c.path,
line: c.line || c.original_line,
body: c.body,
severity: categorizeSeverity(c.body),
url: c.html_url,
}));
return coderabbitComments;
}
function categorizeSeverity(body: string): "critical" | "warning" | "suggestion" {
const lower = body.toLowerCase();
if (lower.includes("security") || lower.includes("vulnerability") || lower.includes("injection")) {
return "critical";
}
if (lower.includes("bug") || lower.includes("error") || lower.includes("issue")) {
return "warning";
}
return "suggestion";
}
Step 3: Post Commands to CodeRabbit via PR Comments // Programmatically trigger CodeRabbit actions
async function sendCodeRabbitCommand(
owner: string, repo: string, prNumber: number, command: string
) {
await octokit.issues.createComment({
owner, repo, issue_number: prNumber,
body: `@coderabbitai ${command}`,
});
}
// Available commands:
// "full review" - Complete review from scratch
// "summary" - Generate walkthrough summary
// "resolve" - Mark all comments resolved
// "generate-docstrings" - Generate docstrings for functions
// "configuration" - Show current config as YAML
// "run <recipe>" - Run a finishing touch recipe
Step 4: Build a Review Dashboard Script #!/bin/bash
# scripts/coderabbit-dashboard.sh - Review metrics for last 50 PRs
set -euo pipefail
ORG="${1:?Usage: $0 <org> <repo>}"
REPO="${2:?Usage: $0 <org> <repo>}"
echo "=== CodeRabbit Review Dashboard ==="
echo "Repository: $ORG/$REPO"
echo ""
# Count PRs with CodeRabbit reviews
TOTAL=$(gh api "repos/$ORG/$REPO/pulls?state=closed&per_page=50" --jq 'length')
REVIEWED=0
for PR_NUM in $(gh api "repos/$ORG/$REPO/pulls?state=closed&per_page=50" --jq '.[].number'); do
HAS_CR=$(gh api "repos/$ORG/$REPO/pulls/$PR_NUM/reviews" \
--jq '[.[] | select(.user.login=="coderabbitai[bot]")] | length' 2>/dev/null || echo "0")
[ "$HAS_CR" -gt 0 ] && REVIEWED=$((REVIEWED + 1))
done
echo "Review Coverage: $REVIEWED/$TOTAL PRs reviewed ($(( REVIEWED * 100 / TOTAL ))%)"
Step 5: GitHub Actions Automation # .github/workflows/coderabbit-gate.yml
# Block merge until CodeRabbit has reviewed
name: CodeRabbit Review Gate
on:
pull_request_review:
types: [submitted]
jobs:
check-coderabbit:
if: github.event.review.user.login == 'coderabbitai[bot]'
runs-on: ubuntu-latest
steps:
- name: Check review state
uses: actions/github-script@v7
with:
script: |
const { data: reviews } = await github.rest.pulls.listReviews({
owner: context.repo.owner,
repo: context.repo.repo,
pull_number: context.issue.number,
});
const crReview = reviews.find(r => r.user.login === 'coderabbitai[bot]');
if (crReview?.state === 'CHANGES_REQUESTED') {
core.setFailed('CodeRabbit requested changes. Address feedback before merging.');
} else {
core.info(`CodeRabbit review state: ${crReview?.state || 'pending'}`);
}
Output
GitHub API integration for fetching CodeRabbit review data
Automated command posting to trigger CodeRabbit actions
Review metrics dashboard script
CI gate that enforces CodeRabbit approval before merge
Error Handling Issue Cause Solution Review not found PR too new Wait 2-5 minutes for review to complete 403 from GitHub API Token missing scopes Ensure repo scope on personal access token Bot login doesn't match Different app slug Check with coderabbitai[bot] (includes [bot] suffix) Rate limited by GitHub Too many API calls Use pagination and caching for bulk queries
Resources
Next Steps Apply patterns in coderabbit-core-workflow-a for real-world review workflows.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in (single or multi-account), or reading/injecting/running secrets via op.
CLI to manage emails via IMAP/SMTP. Use `himalaya` to list, read, write, reply, forward, search, and organize emails from the terminal. Supports multiple accounts and message composition with MML (MIME Meta Language).
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in (single or multi-account), or reading/injecting/running secrets via op.
CLI to manage emails via IMAP/SMTP. Use `himalaya` to list, read, write, reply, forward, search, and organize emails from the terminal. Supports multiple accounts and message composition with MML (MIME Meta Language).