Atomic agents' code-first approach is great for composability, but it also means governance patterns (audit trails, kill switches, human oversight) are left to individual developers to implement. Easy to miss in review.
A CI scanner that automatically checks every PR for governance gaps in agent files would fit well with the code-first philosophy - catch compliance issues the same way you'd catch type errors or test failures.
I've been working on asqav-compliance, a GitHub Action that does this for agent frameworks. Would a recommended CI integration or built-in governance validation be useful for the project?
Atomic agents' code-first approach is great for composability, but it also means governance patterns (audit trails, kill switches, human oversight) are left to individual developers to implement. Easy to miss in review.
A CI scanner that automatically checks every PR for governance gaps in agent files would fit well with the code-first philosophy - catch compliance issues the same way you'd catch type errors or test failures.
I've been working on asqav-compliance, a GitHub Action that does this for agent frameworks. Would a recommended CI integration or built-in governance validation be useful for the project?