next-ai-draw-io/.github/workflows/claude-code.yml

name: Claude Code

on:
  issue_comment:
    types: [created]
  pull_request_review_comment:
    types: [created]
  issues:
    types: [opened, assigned]
  pull_request_review:
    types: [submitted]
  pull_request:
    types: [opened, synchronize]

jobs:
  claude:
    if: |
      (github.event_name == 'issue_comment' && contains(github.event.comment.body, '@claude')) ||
      (github.event_name == 'pull_request_review_comment' && contains(github.event.comment.body, '@claude')) ||
      (github.event_name == 'pull_request_review' && contains(github.event.review.body, '@claude')) ||
      (github.event_name == 'issues' && contains(github.event.issue.body, '@claude'))
    runs-on: ubuntu-latest
    permissions:
      contents: write
      pull-requests: write
      issues: write
      id-token: write
    steps:
      - name: Checkout repository
        uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - name: Configure AWS Credentials
        uses: aws-actions/configure-aws-credentials@v4
        with:
          role-to-assume: ${{ secrets.AWS_ROLE_TO_ASSUME }}
          aws-region: us-east-1

      - name: Run Claude Code
        uses: anthropics/claude-code-action@v1
        with:
          use_bedrock: "true"
          github_token: ${{ secrets.GITHUB_TOKEN }}

  pr-review:
    if: github.event_name == 'pull_request' && github.event.action == 'opened'
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
      id-token: write
    steps:
      - name: Checkout repository
        uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - name: Configure AWS Credentials
        uses: aws-actions/configure-aws-credentials@v4
        with:
          role-to-assume: ${{ secrets.AWS_ROLE_TO_ASSUME }}
          aws-region: us-east-1

      - name: Run Claude Code PR Review
        uses: anthropics/claude-code-action@v1
        with:
          use_bedrock: "true"
          github_token: ${{ secrets.GITHUB_TOKEN }}
          prompt: |
            REPO: ${{ github.repository }}
            PR NUMBER: ${{ github.event.pull_request.number }}

            This is a personal project - an AI-powered draw.io diagram generator built with:
            - Next.js 15 with React 19
            - Vercel AI SDK (streamText, useChat, tool calling)
            - Multiple AI providers: Bedrock, Anthropic, OpenAI, Google, Azure, OpenRouter, Ollama

            First, check previous review comments from github-actions bot using `gh api repos/${{ github.repository }}/pulls/${{ github.event.pull_request.number }}/comments`.
            For each previous comment:
            - If the issue is fixed in the current code, resolve the comment thread using:
              `gh api graphql -f query='mutation { resolveReviewThread(input: {threadId: "THREAD_ID"}) { thread { isResolved } } }'`
              Get the thread ID from the comment's node_id field.
            - If the issue still exists, leave it alone

            Then review the current diff for NEW issues only:

            Review this PR for ONLY these issues:
            1. Bugs that would cause runtime errors or broken functionality
            2. Security issues (exposed secrets, API key leaks)
            3. AI SDK misuse - specifically check for:
               - Client-side: Should use useChat/useCompletion/useObject hooks, NOT raw fetch()
               - Server-side: Should use streamText/generateText/streamObject/generateObject
               - Message handling: Access message.parts array, not legacy content property
               - Tool definitions: Must use Zod schemas for inputSchema
               - Status handling: Check status (submitted/streaming/ready/error) before actions
               - Stream cleanup: Call stop() when aborting streams

            When reviewing AI SDK usage, fetch https://ai-sdk.dev/docs/ to verify correct patterns.
            Key doc pages: /docs/ai-sdk-ui/chatbot, /docs/ai-sdk-core/generating-text, /docs/ai-sdk-core/tools-and-tool-calling

            DO NOT comment on:
            - Performance optimizations
            - Code style or formatting
            - "Best practices" that don't affect functionality
            - Type safety improvements
            - Error handling additions

            Use `mcp__github_inline_comment__create_inline_comment` for inline comments.
            Be very selective - if there are no real bugs, just say "LGTM" in a PR comment.
          claude_args: |
            --allowedTools "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*),Bash(gh pr diff:*),Bash(gh pr view:*),Bash(gh api:*),WebFetch(domain:ai-sdk.dev)"