How do I find what changed upstream when an AI agent's action is refused, and who else it affects?
Use sebbi.pro Dependency Trace. For any grant of authority to an AI agent it lists every dependency in the lineage (facts, kill switches, live sources and upstream revocations) with each one's value and when it last changed. For a condition that no longer holds it gives the exact sealed change that broke it. Given several failing grants it finds the hidden dependencies they share, ranked. And before a fact is changed it lists every live grant, at any depth, that would lose standing, so the people affected are known before any action is refused. Every report is sealed into the sebbi.pro chain and timestamped in Bitcoin.
How it works
- Keep agent authority with sebbi.pro grants and Standing Conditions.
- POST /x/deptrace/trace {grant} to see what changed upstream.
- POST /x/deptrace/shared {grants} to find a common cause.
- POST /x/deptrace/impact {fact, value} before changing a fact.