Skip to main content
All terms
Patterns

Agentic Terminal Coding

Agentic coding measured in a live terminal, where the AI runs real shell commands to complete tasks the way a developer would.

Definition

Agentic terminal coding is a slice of agentic coding where the AI agent works directly in a command-line terminal — running real shell commands, installing tools, editing files, and checking output — to complete a task, rather than just producing code for a human to run. Because a terminal can do almost anything on a computer, it is a demanding test of an agent's ability to plan, use tools, and recover from errors over many steps. Benchmarks like Terminal-Bench and Frontier-Bench measure exactly this, and it appears as a common capability axis on model comparison pages.