Background readingRising
Agentic Coding
noun
Using an AI agent to work through a coding task: read files, make changes, run tests and try again.
Background reading: this term helps explain some of the references in the dictionary, rather than being slang or a meme itself. Browse the AI culture dictionary.
What does agentic coding mean?
You ask an agent to fix a bug. It reads the relevant files, edits the code, runs a test and uses the result to decide what to do next. That sequence is what makes the workflow agentic. The person can still review each change or limit which actions the agent may take. Different tools give the agent different amounts of freedom.
How people use it
I gave the coding agent a failing test and a narrow goal, then reviewed the patch it produced.
Example written for this dictionary.
Here, agentic describes what the tool can do. Vibe coding describes how the person works with it. Someone might watch an agent closely, or accept its changes with barely a glance. The same tool allows both approaches.
Why it matters
The appeal is being able to hand over a task with several steps. The catch is that an early misunderstanding can carry through the whole job. Reviewing the result matters more than whether the agent looked busy getting there.
Where it came from
The phrase comes from the wider discussion of AI agents: systems that take actions toward a goal. We have not identified a first use of the coding term. We include it as background because it helps explain vibe coding.