Hey Product Hunt 👋 I’m excited to launch MaruCheck today. AI coding agents are getting incredibly good at writing software, but there’s a problem I kept running into: The same AI that writes the implementation can also write the tests. If it misunderstood the requirement in the first place, both the code and the tests can share the same mistake. Everything turns green while the product is still wrong. That’s why I built MaruCheck. MaruCheck acts as an independent QA and verification layer for AI-generated software. Instead of only asking, “Do the tests pass?”, it asks: “Does this implementation still match what the product was actually supposed to do?” Some of what MaruCheck currently includes: → Quality Contracts for defining intended product behavior → Git diff and change-impact analysis → Risk-based verification → Automated test orchestration → Semantic drift detection → QA Memory for previous bugs and regressions → MCP integration for AI coding agents → CLI workflows → GitHub / CI verification → Evidence-based release checks One of my favorite parts is semantic drift protection. If the approved requirement says: Free users can upload 5 files. and an AI-generated change silently changes that to 10, MaruCheck doesn’t just rewrite the test and move on. It flags the change and asks whether the product behavior was intentionally changed. The name MaruCheck is inspired by the Kobayashi Maru concept: testing what happens when things don’t go according to the expected happy path. I’ve decided to make the project open source because I’d really like other developers to help shape it. I’d love for you to: ⭐ Star the repository 🛠 Try MaruCheck on a real project 🐛 Report bugs 💡 Suggest features 🤝 Contribute code, integrations, tests, docs, or ideas 🔁 Share it with developers using AI coding tools Website: https://marucheck.dev Core repository: https://github.com/Kidus-M/MaruC... Website repository: https://github.com/Kidus-M/MaruC... I’d especially love feedback from people building regularly with Codex, Claude Code, Cursor, Windsurf, and other coding agents. The coding agent builds. MaruCheck proves.

MaruCheck

Test what your AI didn't.

Test what your AI didn't.