RunasSpec

RunasSpec is a proposed specification for describing run-as policies in computing environments. It provides a formal schema for expressing how and under what identities a process may execute commands or programs on behalf of another user. The goal is to enable consistent policy definition, auditing, and cross-system interoperability in privilege-management frameworks across operating systems, containers, and orchestration platforms.

The specification defines a policy document that includes subject identifiers (the initiating user or service), target identities (the impersonated user), allowed actions (execute, read, write), scope (system, application, container), constraints (time windows, source endpoints, environment restrictions), and auditing metadata. It supports declarative rules that can be evaluated by policy engines embedded in platforms or by central authorization services. Common features include inheritance of permissions, whitelisting versus blacklisting, and mechanisms for fallback or default rules.

Syntax and formats are described as machine-readable schemas with preferred encodings in JSON and YAML, plus optional bindings to existing access-control models such as RBAC or ABAC. The design emphasizes security considerations, including least-privilege enforcement, robust auditing, and tamper-evident policy storage.

History and status: RunasSpec has appeared in security literature and industry forums as a conceptual standard; several vendors have prototyped implementations, but there is no universal ratification. See also: run as, sudo, privilege escalation, policy-as-code.