In this short update video, I share some developments made to the Test-Driven Agentic Behaviours test framework demoed before. Specifically, the way it uses subagents and in particular how the framework tests itself.
I am evolving this into an open source framework under the name ‘stagentic-ai’.
Stage + Agentic: because each test run is a rehearsal, evaluated, and you improve your directions (via the SKILL.md file you’re developing) until you get the agent, or actor’s, desired ‘performance’.
UPDATE: There was a serious issue, hiding in plain sight, in this demo. Did you spot it? My tests did…
Watch the previous video here:
Read the background article to this here:
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.