Automated Code Review in CI/CD

What the Code Review Agent Does

Unlike static linters or rule-based scanners, CurieTech AI's Code Review Agent uses intent-driven analysis to:

• Catch subtle bugs and logic flaws
• Flag anti-patterns and inefficient design choices
• Recommend improvements aligned with your architecture
• Enforce security and performance best practices
• Explain rationale with context-specific suggestions

This agent functions like a senior reviewer embedded in your team — thorough, consistent, fast, and always aligned with your development patterns.

---

If you're using your own script, you can trigger the review using:

curl -X POST "https://platform.curietech.ai/api/code-review/run" \
  -H "Authorization: Bearer $CURIE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "conn_type": "One of: github/bitbucket/gitlab/azure_devops",
    "org_name": "Name of the organization",
    "repo_name": "Name of the repository without the .git suffix",
    "branch_name": "Name of the feature branch",
    "pr_number": [Integer] PR number,
    "prompt": "[Optional] Any additional instructions for the code review agent"
  }'

Or if you want to use it in CI/CD for dynamic values, you can use the below steps.

Note: For CI/CD steps, the org_name in the JSON payload should be projectname/orgname. The project name has to be hardcoded.

---

GitHub Actions

Steps

1. Open your repository in GitHub.
2. Go to Settings → Secrets and variables → Actions and click New repository secret.
   • Name: CURIE_API_KEY
   • Value: Your CurieTech API key.
3. Create .github/workflows/curie-code-review.yml in your repo.
4. Commit and push. This will run on every Pull Request.

# This is a basic workflow to help you get started with Actions

name: CI # you can provide any name

# Controls when the workflow will run
on:
  # Triggers the workflow on push or pull request events but only for the "main" branch
  push:
    branches: [ "main" ]
  pull_request:
    branches: [ "main" ]

  # Allows you to run this workflow manually from the Actions tab
  workflow_dispatch:

# A workflow run is made up of one or more jobs that can run sequentially or in parallel
jobs:
  # This workflow contains a single job called "build"
  build:
    # The type of runner that the job will run on
    runs-on: ubuntu-latest

    # Steps represent a sequence of tasks that will be executed as part of the job
    steps:
      # Checks-out your repository under $GITHUB_WORKSPACE, so your job can access it
      - uses: actions/checkout@v4

      # Runs a single command using the runners shell
      - name: Curie-review
        run: |
          curl -X POST "https://platform.curietech.ai/api/code-review/run" \
            -H "Authorization: Bearer ${{ secrets.CURIE_API_KEY }}" \
            -H "Content-Type: application/json" \
            -d '{
              "conn_type": "github",
              "org_name": "org_name", # provide your org name here
              "repo_name": "'"${GITHUB_REPOSITORY#*/}"'",
              "branch_name": "'"${{ github.head_ref || github.ref_name }}"'",
              "pr_number": ${{ github.event.pull_request.number }},
              "prompt": "CI-triggered code review" # you can change the prompt according to your wish
            }'

---

Azure DevOps

Steps

1. Create a pipeline for your selected repo.
2. Paste the YAML below into your pipeline.
3. Save, then go back to the same pipeline → Edit → Variables:
   • Name: CURIE_API_KEY (mark as secret)
   • Value: Your CurieTech API key.
4. Run the pipeline on a Pull Request.

# azure-pipelines.yml
trigger: none
pr: none

pool:
  vmImage: 'ubuntu-latest' # or 'windows-latest'/'macos-latest'

steps:
- checkout: self

- task: PowerShell@2
  displayName: 'Run Curie AI Code Review'
  env:
    CURIE_API_KEY: $(CURIE_API_KEY)
  inputs:
    targetType: 'inline'
    script: |
      if ("$(Build.Reason)" -eq "PullRequest") {
          $prNumber = "$(System.PullRequest.PullRequestId)"
          $sourceBranchName = "$(System.PullRequest.SourceBranch)" -replace 'refs/heads/', ''
          $repoName = "$(Build.Repository.Name)"
          $orgName = "$(System.CollectionUri)".Split('/')[3] # Format: projectname/orgname

          Write-Host "PR Number: $prNumber"
          Write-Host "Source Branch: $sourceBranchName"
          Write-Host "Repository Name: $repoName"
          Write-Host "Organization Name: $orgName"

          $payload = @{
              conn_type = "azure_devops"
              org_name = $orgName
              repo_name = $repoName
              branch_name = $sourceBranchName
              pr_number = [int]$prNumber
          } | ConvertTo-Json -Compress

          Write-Host "Sending payload: $payload"

          Invoke-RestMethod -Uri "https://platform.curietech.ai/api/code-review/run" `
              -Method POST `
              -Headers @{
                  "Authorization" = "Bearer $env:CURIE_API_KEY"
                  "Content-Type" = "application/json"
              } `
              -Body $payload
      } else {
          Write-Host "This pipeline run is not a Pull Request. Skipping Curie AI Code Review."
      }

---

GitLab

Steps

1. Go to your project → Settings → CI/CD → Variables.
   • Name: CURIE_API_KEY (mark as masked and protected).
   • Value: Your CurieTech API key.
2. Commit .gitlab-ci.yml to your repository.

# .gitlab-ci.yml
stages:
  - code_review

curie_ai_code_review:
  stage: code_review
  image: alpine/curl:latest # A lightweight image with curl installed, or use your preferred image
  rules:
    - if: $CI_MERGE_REQUEST_IID # This rule ensures the job only runs for merge requests
      when: always # Run always if it's a merge request, otherwise skip

  script:
    - |
      # Extract branch name, removing 'refs/heads/' prefix if present
      BRANCH_NAME="${CI_MERGE_REQUEST_SOURCE_BRANCH_NAME}"

      # CI_PROJECT_NAMESPACE gives the group path (e.g., 'my-group/my-subgroup')
      # CI_PROJECT_NAME gives the project name
      # For 'org_name', you typically want the top-level group
      # If your project is 'my-group/my-subgroup/my-project', CI_PROJECT_NAMESPACE is 'my-group/my-subgroup'
      # You might need to adjust 'org_name' based on how you define 'organization' in your GitLab structure.
      # For a simple top-level group, CI_PROJECT_NAMESPACE should suffice.
      # If you need just the root group name:
      ORG_NAME=$(echo "$CI_PROJECT_NAMESPACE" | cut -d'/' -f1) # Format: projectname/orgname

      echo "Merge Request ID: ${CI_MERGE_REQUEST_IID}"
      echo "Source Branch Name: ${BRANCH_NAME}"
      echo "Project Name: ${CI_PROJECT_NAME}"
      echo "Organization/Group Name: ${ORG_NAME}"

      PAYLOAD='{
        "conn_type": "gitlab",
        "org_name": "'"$ORG_NAME"'",
        "repo_name": "'"${CI_PROJECT_NAME}"'",
        "branch_name": "'"$BRANCH_NAME"'",
        "pr_number": '"${CI_MERGE_REQUEST_IID}"'
      }'

      echo "Sending payload: $PAYLOAD"

      curl -X POST "https://platform.curietech.ai/api/code-review/run" \
        -H "Authorization: Bearer ${CURIE_API_KEY}" \
        -H "Content-Type: application/json" \
        -d "$PAYLOAD"

  variables:
    # Ensure this variable is masked and protected in GitLab CI/CD settings
    CURIE_API_KEY: $CURIE_API_KEY

---

Bitbucket

Steps

1. Go to Repository settings → Repository variables.
   • Name: CURIE_API_KEY
   • Value: Your CurieTech API key.
2. Commit bitbucket-pipelines.yml to your repository.

# bitbucket-pipelines.yml
image: atlassian/default-image:2 # You can use a different image if curl isn't available

pipelines:
  pull-requests:
    '**': # This pattern means the pipeline will run for all pull requests
      - step:
          name: Curie AI Code Review
          script:
            # Construct the JSON payload
            - >
              PAYLOAD='{
                "conn_type": "bitbucket",
                "org_name": "${BITBUCKET_REPO_OWNER}", # Format: projectname/orgname
                "repo_name": "${BITBUCKET_REPO_SLUG}",
                "branch_name": "${BITBUCKET_BRANCH}",
                "pr_number": ${BITBUCKET_PR_ID}
              }'
            - echo "Sending payload: $PAYLOAD"
            - >
              curl -X POST "https://platform.curietech.ai/api/code-review/run" \
              -H "Authorization: Bearer ${CURIE_API_KEY}" \
              -H "Content-Type: application/json" \
              -d "$PAYLOAD"

---

By following the setup instructions for your chosen CI/CD platform, you can seamlessly integrate CurieTech AI’s Code Review Agent into your development workflow. This ensures consistent, high-quality, and secure code reviews — automated directly within your pull request process — saving time and improving code quality across your team.