Skip Review Labeler

^{Image generated by Nano Banana Pro.}

AI-powered GitHub Action that automatically labels low-risk pull requests for skipping human code review.

Why

Code review is essential, but not all changes carry the same risk. Typo fixes, i18n updates, and formatting changes don't need the same scrutiny as business logic changes.

The Problem:

• Reviewers waste cognitive energy on trivial PRs
• Simple fixes get stuck waiting for review
• Context switching between complex and trivial PRs reduces focus

The Solution:

• Automatically identify low-risk PRs with AI analysis
• Reduce review noise by ~20% (based on real-world usage)
• Enable zero-latency deployment for trivial changes
• Let reviewers focus on high-impact, complex changes

Features

• AI-Powered Analysis - Uses OpenAI (or compatible APIs) to analyze PR changes
• Conservative by Default - Only labels PRs with high confidence scores
• Transparent Decisions - Adds explanatory comments to labeled PRs
• Configurable - Customize model, threshold, label name, and more
• Skip-Review Categories - Detects typos, i18n updates, UI tweaks, formatting, unused-code cleanup, and safe dependency bumps

How It Works

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│   PR Opened/    │────▶│  Fetch Diff &   │────▶│  AI Analyzes    │
│   Synchronized  │     │  File Changes   │     │  Changes        │
└─────────────────┘     └─────────────────┘     └────────┬────────┘
                                                         │
                        ┌─────────────────┐              │
                        │  Add Label &    │◀─── Yes ─────┤ Eligible &
                        │  Comment        │              │ High Confidence?
                        └─────────────────┘              │
                                                         │
                        ┌─────────────────┐              │
                        │  Skip (No       │◀─── No ──────┘
                        │  Action)        │
                        └─────────────────┘

Quick Start

1. Add OpenAI API Key

Add your OpenAI API key as a repository secret:

1. Go to Settings > Secrets and variables > Actions
2. Click New repository secret
3. Name: OPENAI_API_KEY
4. Value: Your OpenAI API key

2. Create Workflow File

Create .github/workflows/skip-review-labeler.yml:

name: AI Skip-Review Labeler

on:
  pull_request:
    types: [opened, synchronize, reopened]

concurrency:
  group: ${{ github.workflow }}-${{ github.event.pull_request.number }}
  cancel-in-progress: true

jobs:
  analyze:
    runs-on: ubuntu-latest
    # Skip if already labeled or opened by a bot account
    if: |
      !contains(github.event.pull_request.labels.*.name, 'skip-review') &&
      github.event.pull_request.user.type == 'User'  # Exclude bot-created PRs

    permissions:
      contents: read
      pull-requests: write

    steps:
      - uses: chatbotgang/skip-review-labeler@v1
        with:
          github_token: ${{ secrets.GITHUB_TOKEN }}
          openai_api_key: ${{ secrets.OPENAI_API_KEY }}

Configuration

Input	Description	Required	Default
github_token	GitHub token for API access	Yes	-
openai_api_key	OpenAI API key (or compatible)	Yes	-
model	AI model to use	No	gpt-5-mini
confidence_threshold	Minimum confidence to apply label (0-100)	No	80
label_name	Label to apply when eligible	No	skip-review
max_diff_size	Maximum diff size in characters	No	50000
add_comment	Add explanatory comment to PR	No	true

Example with All Options

- uses: chatbotgang/skip-review-labeler@v1
  with:
    github_token: ${{ secrets.GITHUB_TOKEN }}
    openai_api_key: ${{ secrets.OPENAI_API_KEY }}
    model: gpt-4o
    confidence_threshold: 90
    label_name: auto-merge-eligible
    max_diff_size: 100000
    add_comment: true

Skip-Review Categories

The action identifies the following low-risk change types:

1. Fix Typos

Spelling or grammar corrections in comments, documentation, or string literals that don't affect functionality.

- // Calcualte the total price
+ // Calculate the total price

2. Update i18n Key

Changes to internationalization files, translation strings, or translation key references.

- "campaigns.title": "Campaigns"
+ "campaigns.title": "Marketing Campaigns"

3. Update UI Style

Visual-only changes to CSS, styled-components, or inline styles without logic changes.

- padding: 12px 16px;
+ padding: 16px 24px;

4. Code Formatting

Automated formatting changes from tools like Prettier or ESLint.

- const sum=a+b;
+ const sum = a + b;

5. Remove Unused Code

Cutting deprecated endpoints, feature flags, or configs that have zero remaining callers.

-// Temporary shim until everyone hits v2
-app.use('/api/v1/members', legacyMembersRouter);
 app.use('/api/v2/members', membersRouter);

6. Safe Dependency Version Bump

Patch or minor version updates that only touch dependency manifests or lockfiles.

-  "typescript": "5.5.4"
+  "typescript": "5.5.5"

Outputs

Output	Description
eligible	Whether the PR is eligible for skip-review (true/false)
confidence	AI confidence score (0-100)
category	Detected category or none
reasoning	AI reasoning for the decision

The action also automatically outputs a summary to GITHUB_STEP_SUMMARY.

Using Outputs

- uses: chatbotgang/skip-review-labeler@v1
  id: labeler
  with:
    github_token: ${{ secrets.GITHUB_TOKEN }}
    openai_api_key: ${{ secrets.OPENAI_API_KEY }}

- name: Check result
  run: |
    echo "Eligible: ${{ steps.labeler.outputs.eligible }}"
    echo "Confidence: ${{ steps.labeler.outputs.confidence }}%"
    echo "Category: ${{ steps.labeler.outputs.category }}"
    echo "Reasoning: ${{ steps.labeler.outputs.reasoning }}"

Security Considerations

• Never commit API keys - Always use repository secrets
• Review the label - The skip-review label is a suggestion, not a mandate
• Conservative defaults - 80% confidence threshold ensures high accuracy
• Audit trail - Comments explain why each PR was labeled
• Human override - Remove the label manually if you disagree

What is NOT Eligible

The AI is trained to reject these change types:

• Any logic or functional changes
• API endpoint modifications
• Configuration file changes
• Dependency updates
• Test file changes
• Security-related code
• Database queries or migrations

Troubleshooting

403 Permission Error

Ensure your workflow has the required permissions:

permissions:
  contents: read
  pull-requests: write

Label Not Applied

Check the workflow run logs. Common reasons:

1. Confidence too low - The AI wasn't confident enough (below threshold)
2. Not eligible - Changes include logic modifications
3. Already labeled - PR already has the skip-review label
4. Diff too large - Exceeds max_diff_size

API Rate Limits

If hitting OpenAI rate limits:

1. Use a model with higher rate limits
2. Add concurrency controls to your workflow
3. Consider using Azure OpenAI endpoint

Using with Azure OpenAI

The action supports Azure OpenAI by configuring the API endpoint:

- uses: chatbotgang/skip-review-labeler@v1
  with:
    github_token: ${{ secrets.GITHUB_TOKEN }}
    openai_api_key: ${{ secrets.AZURE_OPENAI_KEY }}
    openai_base_url: https://your-resource.openai.azure.com
    model: your-deployment-name

Cost Estimation

Using gpt-5-mini (default):

• Average PR diff: ~2,000 tokens input, ~200 tokens output
• Cost per PR: ~$0.0003
• 1,000 PRs: ~$0.30

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Submit a pull request

License

Apache License 2.0 - see LICENSE for details.