Authored injections of context that are resolved at authentication time

摘要

Techniques are described for enabling principals to inject context information into a credential (e.g. session credential). Once the credential has been issued, any arbitrary principal is allowed to inject context information into the existing credential. The injected context is scoped to the principal that made the injection. Subsequently, at authentication time, when the credential is used to request access to a particular resource, the system can verify whether the principal that made the injection is trusted and if the principal is deemed trusted, the context information can be applied to a policy that controls access to one or more resources, or can alternatively be translated into some context residing in a different namespace which can then be applied to the policy. In addition, the system enables arbitrary users to insert additional deny statements into an existing credential, which further restrict the scope of permissions granted by the credential.