AI Pull Request Review Bot

1. Create a GitHub webhook (Settings > Webhooks) pointing to your automation platform's webhook URL. Select the "Pull requests" event with action "opened" and "synchronize". 2. Add a GitHub node to fetch the PR diff using the GitHub API: GET /repos/{owner}/{repo}/pulls/{pull_number}/files. This returns the changed files with patch content. 3. Add a Code/Function node to format the diff into a concise prompt. Limit the diff to ~3,000 tokens per file to stay within Claude's context — truncate large files and note the truncation. 4. Add an HTTP Request node to call the Claude API (POST https://api.anthropic.com/v1/messages). Set headers: x-api-key to your Anthropic API key, anthropic-version to 2023-06-01. Use model "claude-sonnet-4-20250514" with max_tokens 1024. Craft a system prompt that instructs Claude to act as a senior code reviewer: summarize changes, flag bugs, security issues, and performance concerns. 5. Parse Claude's response from the JSON body (content[0].text) and format it into a readable Slack message with sections for Summary, Issues Found, and Suggestions. 6. Add a Slack node to post the review to your #code-reviews channel. Include the PR title, author, link, and Claude's analysis. 7. Optionally add a GitHub node to post Claude's review as an actual PR comment using POST /repos/{owner}/{repo}/issues/{pull_number}/comments. 8. Test by opening a small PR and verifying the review appears in Slack within 30 seconds.