Show HN: Open Agent Spec. Treat AI agents like typed functions not prompt chains

2 comments

• $andrewvector 5 hours ago

  Curious how others here are handling structure / validation in their agent workflows, especially around output reliability.

  I've been experimenting more in my dev sandbox here - https://www.openagentspec.dev/

• $andrewvector 6 hours ago

  [dead]