Overview

The agent is CodeNeter's coding assistant. Give it a task, such as a new feature, a bug fix, or a question about how something works, and it carries the work out in your project. It searches and reads the code, edits files, and runs terminal commands along the way.

What it can do

• Find its way around the code. It lists folders, searches files, and reads the parts that matter for the task.
• Change files. It creates, edits, and deletes files. Each change appears in the conversation, and the Changes tab shows the full diff.
• Run commands. It runs terminal commands in the project folder, such as installs, tests, or a build, and reads their output.
• Use the browser. In the desktop app's built-in browser, it can load pages, interact with them, take screenshots, and read console errors.
• Look things up. It can search the web, fetch pages, and use tools from MCP servers you've connected.

Some actions can wait for your approval before they run, such as terminal commands. Your approval settings decide which ones.

Give it context

The agent does better work when it knows the outcome you want, the part of the code involved, and anything that shouldn't change. Mention files or folders with @, and attach a screenshot when a picture explains the problem, if the selected model accepts images. For UI changes, you can point at elements on the page instead; see Design with the agent.

Work alongside it

Messages you send while the agent is working wait in a queue and go out when the current run ends. Send now stops the run and sends that message right away. The stop button ends a run without sending anything.

Choose how it works

Agent mode makes changes directly. Plan works out an approach with you first, and Ask answers questions without editing. See Planning and modes.