SHRIKEDelegation chain
A delegation chain is the tree of agents created when an AI agent spawns sub-agents for sub-tasks, which may spawn their own. Each hop can widen access and blur accountability, so security controls must track the full chain: which root agent a given action ultimately serves, and at what depth.
Delegation is how agent systems scale, and how permissions quietly spread. A sub-agent typically inherits capabilities from its parent; three hops later, a task nobody reviewed is running with access granted for a different purpose. Governing chains means propagating identity and scope down the tree, capping depth, and treating high-consequence actions as review points regardless of where in the tree they occur.
Related
Shrike governs AI agent actions in real time: every command, query, and tool call evaluated against policy before it executes. Start with what is action governance or the quickstart.