Skip to content

Comment on Provide agents with automated feedback

Comments

I've been slowly working on https://blocksai.dev/ which is a framework for building feedback loops for agentic coding purposes. It just exposes a CLI that can run custom validators against anything with a spec in the middle. It's goal being like the blog post is to make sure their is always a feedback loop for the agent, be it programmatic test, semantic linting, visual outputs, anything!

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.