AI Agent Security (Course)
Agent Loop
Also: agentic loop
The repeating cycle in which a harness shows context to a model, receives a tool call, runs it and feeds the result back.
In more detail
A coding agent is not a single question and answer. The harness assembles the context, the model replies with an answer or a tool call, the harness decides whether to allow the call, runs it, appends the result to the context and repeats until the model reports it is finished. Each pass is a chance for untrusted text to enter and a chance for a gate to deny an action, which is why the loop is the map for most agent attacks and defenses.