The scanner-comparison.yml workflow provides a comprehensive benchmark comparing multiple secret scanning tools on the OWASP WrongSecrets repository. This helps understand the relative effectiveness of different secret detection tools.
The workflow currently tests 7 different secret scanning tools:
- TruffleHog - Docker-based secret scanner with verified results
- git-secrets - AWS Labs' git hook for preventing secrets in commits
- gitleaks - High-performance secret scanner with JSON output
- detect-secrets - Yelp's enterprise secret scanner with baseline functionality
- gittyleaks - Python-based secret scanner with broad pattern detection
- whispers - Skyscanner's structured secret detection tool
- trufflehog3 - Python version of TruffleHog with additional features
# Trigger manually via GitHub Actions UI
# Go to Actions > Secret Scanner Comparison Benchmark > Run workflowThe workflow automatically runs every Sunday at 02:00 UTC to provide regular benchmarking.
The workflow generates a summary report showing:
| Scanner | Secrets Found |
|---------|---------------|
| TruffleHog | X |
| git-secrets | X |
| gitleaks | X |
| detect-secrets | X |
| gittyleaks | X |
| whispers | X |
| trufflehog3 | X |
Based on manual testing, the tools typically find:
- detect-secrets: ~98 potential secrets
- gitleaks: ~106 secrets
- gittyleaks: ~137 findings
Results may vary as the repository evolves and tools update their detection patterns.
The workflow includes robust error handling:
- Tools that fail to install will show 0 results
- Network timeouts are handled gracefully
- Each scanner job runs independently in parallel
- Failed scans don't block the summary report
- git-secrets: Compiled from source (no official GitHub Action)
- gitleaks: Downloaded as binary from GitHub releases
- Python tools: Installed via pip with timeout handling
- TruffleHog: Uses official Docker image
Different tools have varying approaches:
- Some focus on verified/confirmed secrets
- Others detect potential patterns that may be false positives
- Count differences are expected and help understand tool characteristics
This benchmark helps users choose appropriate tools for their security scanning needs based on detection coverage and accuracy requirements.