SwarmAI is an automated testing system that uses multiple AI agents working in parallel to:

• Find security vulnerabilities (XSS, SQL injection, CSRF, etc.)
• Detect performance bottlenecks
• Identify real-world usage issues
• Test edge cases and error handling
• Validate API endpoints and data flows

When you start a SwarmAI run:

1. Agent Initialization: Multiple AI agents are spawned, each with a specific testing focus
2. Parallel Testing: Agents work simultaneously to maximize coverage
3. Pattern Detection: Agents identify vulnerabilities, performance issues, and bugs
4. Report Generation: Results are compiled into actionable reports
5. Auto-Patch Suggestions: Fixes are automatically generated for detected issues

SwarmAI uses a headless browser (Playwright + Chromium) that runs on the Sublyzer backend, not inside your SDK code.

• No extra browser libraries are required in your app or SDK integration.
• On Sublyzer Cloud and the official backend image, Playwright/Chromium is already installed and wired into TestingAI/SwarmAI.
• If you self-host a custom backend, just keep the default Dockerfile / npm install flow so the headless engine is provisioned automatically.

To run SwarmAI tests:

1. Ensure your GitHub integration is connected
2. Go to your Integration dashboard → Swarm AI tab
3. Select the test type (Security, Performance, or Full)
4. Configure test depth and number of agents
5. Click "Start SwarmAI Run"

Test types:

• Security: Focuses on finding vulnerabilities and security issues
• Performance: Analyzes performance bottlenecks and optimization opportunities
• Full: Comprehensive testing covering all areas

SwarmAI results include:

• Vulnerabilities Found: List of security issues with severity ratings
• Performance Issues: Slow endpoints, inefficient queries, memory leaks
• Error Patterns: Common error scenarios and edge cases
• Recommendations: Suggested fixes and improvements
• Auto-Patch Files: Ready-to-apply patches for detected issues

• Run SwarmAI tests regularly (weekly or before major releases)
• Start with Security tests for critical applications
• Review results carefully before applying Auto-Patch fixes
• Use Full tests for comprehensive coverage before production deployments
• Monitor test duration and adjust agent count based on codebase size

SwarmAI is designed to complement, not replace, manual testing:

• May produce false positives that require manual review
• Complex business logic may require human verification
• Test coverage depends on codebase structure and documentation
• Some edge cases may only be found through real-world usage