In Decentralized Trust Management (DTM) authorization decisions are made by multiple principals who can also delegate decisions
to each other. Therefore, a policy change of one principal will often affect who gets authorized by another principal. In
such a system of influenceable authorization a number of principals may want to coordinate their policies to achieve long
time guarantees on a set of safety goals.
The problem we tackle in this paper is to find minimal restrictions to the policies of a set of principals that achieve their
safety goals. This will enable building useful DTM systems that are safe by design, simply by relying on the policy restrictions
of the collaborating principals. To this end we will model DTM safety problems in Scoll [1], an approach that proved useful
to model confinement in object capability systems [2].
This work has been supported in part by European Commission FP7 TAS3 project, nr. 216287 and the BSIK project Poseidon.