isagawa-co

3 ms·

1. 1.▲
   by isagawa-co 8mo ago

   Interesting approach to planning via extensions. I took a similar direction with enforcement. A governance loop that hooks into the agent's tool calls and blocks execution until protocol is followed. Every 10 actions (configurable
2. 2.▲
   Open-source AI execution management for test automation(github.com)

   3 points by isagawa-co 8mo ago|2 comments
3. 3.▲
   by isagawa-co 8mo ago

   Hi HN, I built this to solve a problem I kept hitting: AI agents can generate test scripts fast, but without enforcement they produce inconsistent output, skip patterns, and repeat the same mistakes. The platform uses a 5-layer test archite