Architecture guardrails for AI coding agents
No updates yet.
Coding agents can write working code while quietly degrading architecture. I built Guard to check what can be verified directly from the repository, without another LLM.
