Autonomous Multi-agent Debugging Pipeline

Hi, I can build your autonomous multi-agent debugging pipeline for Vendure within 7 days.
​Technical Approach & Tools:
​Orchestration: CrewAI / OpenHands (100% open-source, zero extra costs) to manage agent logic, task planning, and bounded retries.
​Environment & Stack: Node.js, TypeScript, NestJS, GraphQL, TypeORM, PostgreSQL, Redis, BullMQ in isolated Docker containers.
​Verification: Playwright for browser E2E, k6 for load testing, and OWASP ZAP for authorized security checks.
​Answers to Your Specific Questions:
​Failure Evidence & Resource Control: Logs are streamed to a capped 50MB ephemeral buffer with auto-redaction of secrets. On E2E failure, Playwright traces, final screenshots, and console logs are saved to an isolated folder before triggering a safe stop.
​Immutable Acceptance Truth: PostgreSQL/TypeORM database fixtures, E2E specs, and SHA-256 hashed test/security reports serve as the immutable truth.
​Risk Mitigation: Database state is reset using deterministic fixtures before each run to prevent flaky tests, and strict CPU/RAM limits are set in Docker to prevent system overruns.