kirill89/reviewcerberus - Docker Image

2 min read Original article ↗

⁠ReviewCerberus

AI-powered code review tool that analyzes git branch differences and generates comprehensive review reports with structured output.

⁠Quick Start

docker run --rm -it -v $(pwd):/repo \
  -e MODEL_PROVIDER=anthropic \
  -e ANTHROPIC_API_KEY=sk-ant-your-api-key \
  kirill89/reviewcerberus:latest \
  --repo-path /repo --output /repo/review.md

That's it! The review will be saved to review.md in your current directory.

⁠Key Features

  • Comprehensive Reviews: Detailed analysis of logic, security, performance, and code quality
  • Structured Output: Issues organized by severity with summary table
  • Multi-Provider: AWS Bedrock, Anthropic API, Ollama, or Moonshot
  • Smart Analysis: Context provided upfront with prompt caching
  • Git Integration: Works with any repository, supports commit hashes
  • Verification Mode: Experimental Chain-of-Verification⁠ to reduce false positives

⁠Usage Examples

Custom target branch:

docker run --rm -it -v $(pwd):/repo \
  -e MODEL_PROVIDER=anthropic \
  -e ANTHROPIC_API_KEY=sk-ant-your-api-key \
  kirill89/reviewcerberus:latest \
  --repo-path /repo --target-branch develop --output /repo/review.md

With custom review guidelines:

docker run --rm -it -v $(pwd):/repo \
  -e MODEL_PROVIDER=anthropic \
  -e ANTHROPIC_API_KEY=sk-ant-your-api-key \
  kirill89/reviewcerberus:latest \
  --repo-path /repo --instructions /repo/guidelines.md --output /repo/review.md

⁠What's Included

⁠Comprehensive Code Review

Detailed analysis covering:

  • Logic & Correctness: Bugs, edge cases, error handling
  • Security: OWASP issues, access control, input validation
  • Performance: N+1 queries, bottlenecks, scalability
  • Code Quality: Duplication, complexity, maintainability
  • Side Effects: Impact on other system parts
  • Testing: Coverage gaps, missing test cases
  • Documentation: Missing or outdated docs
⁠Structured Output

Every review includes:

  • Summary: High-level overview of changes and risky areas
  • Issues Table: All issues at a glance with severity indicators (🔴 CRITICAL, 🟠 HIGH, 🟡 MEDIUM, 🟢 LOW)
  • Detailed Issues: Each issue with explanation, location, and suggested fix

⁠Configuration

⁠Anthropic API
-e MODEL_PROVIDER=anthropic
-e ANTHROPIC_API_KEY=sk-ant-your-api-key
-e MODEL_NAME=claude-opus-4-5-20251101  # optional
⁠AWS Bedrock (default)
-e AWS_ACCESS_KEY_ID=your_key
-e AWS_SECRET_ACCESS_KEY=your_secret
-e AWS_REGION_NAME=us-east-1
-e MODEL_NAME=us.anthropic.claude-opus-4-5-20251101-v1:0  # optional
⁠Ollama (local models)
-e MODEL_PROVIDER=ollama
-e OLLAMA_BASE_URL=http://host.docker.internal:11434
-e MODEL_NAME=deepseek-v3.1:671b-cloud  # optional
⁠Moonshot
-e MODEL_PROVIDER=moonshot
-e MOONSHOT_API_KEY=sk-your-api-key
-e MOONSHOT_API_BASE=https://api.moonshot.ai/v1  # optional
-e MODEL_NAME=kimi-k2.5  # optional

⁠Command-Line Options

  • --target-branch: Branch to compare against - default: main
  • --output: Output file path or directory
  • --repo-path: Path to git repository - default: /repo
  • --instructions: Path to markdown file with custom review guidelines
  • --verify: Enable verification mode to reduce false positives (experimental)
  • --sast: Enable OpenGrep SAST pre-scan (experimental)
  • --json: Output review as JSON instead of markdown

⁠Requirements

  • Git repository mounted to /repo
  • Either Anthropic API key or AWS Bedrock credentials
  • Output directory must be writable

⁠Links

⁠License

MIT