AFK stands for away from keyboard: running an agent unattended for long stretches while you do something else. You kick off a task, walk away, and let the agent grind through it in agent mode without approving each step.
What makes it safe (or not)
AFK is where agents start to feel like real leverage: hours of work happening while you sleep. It is also where they are most dangerous, because there is no human in the loop to catch a wrong turn early. The whole thing rests on guardrails standing in for the attention you have removed:
- [Automated checks](/ai-coding-dictionary/automated-check): tests, typecheck, lint, and build that the agent must pass before its work counts.
- A sandbox so a mistake cannot reach anything you care about.
- A tight, well-specified task so the agent has less room to wander.
Trust is earned per task
Autonomy is a graduation, not a default. The move is to start with a human in the loop, watch where the agent goes wrong, and only remove yourself once your checks reliably catch those failures. An agent left alone without gates will happily spend an hour producing confident, broken work, and you will not know until you come back.
Related terms
Human in the loop
Human in the loop means keeping a person in the agent's decision path to approve, steer, or verify its work. It is the deliberate counterweight to full autonomy.
Read definition →ConceptAgent mode
Agent mode is a setting where the model runs the loop autonomously, planning and acting on its own, rather than giving a single chat reply or edit. More capable, and it needs more trust.
Read definition →ConceptAutomated check
An automated check is a machine-verifiable gate that agent output has to pass, like tests, a type check, a linter, or a build. It either passes or fails with no judgment, which makes it the backbone of trusting agent code, especially when you are running unattended.
Read definition →AntipatternProvenLethal trifecta
The lethal trifecta is an agent having access to private data, exposure to untrusted content, and a way to communicate outward, all at once. Any two are usually survivable; all three make data theft a matter of someone asking.
Read definition →AntipatternEmergingAgent trap
An agent trap is content placed where an agent will find it, designed to work on an agent rather than a human reader. The target is not a person browsing but an automated reader that acts on what it reads.
Read definition →ConceptBlast radius
Blast radius is how much damage an action can do if it turns out to be wrong. It is the measure that lets you scale review and permissions to consequence instead of treating every change the same.
Read definition →