Show HN: Find where multi-agent AI systems break before production

Category: devtools

Tags: multi-agent, testing, reliability

Score: 7.5/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 7)

swarm-test statically analyzes multi-agent AI system topologies to detect cascade failures, single points of failure, and other reliability issues before deployment, without requiring live LLM calls. It offers a unique approach to testing agent connectivity and resilience, with a comprehensive dashboard and CI integration.

Target audience: AI engineers, backend devs, devops

Repository: https://github.com/surajkumar811/swarm-test · Python · MIT · 2 stars

View on Hacker News