Last updated 13 October 2026
Assign names to Claude Code session tiles
The struggle with anonymous terminal windows
When you run multiple Claude Code sessions at once, the terminal windows start to blend together. You might be working on a bug fix in one tab and a new feature in another. Without a way to label these sessions, you end up clicking through every open window to find your place. This confusion creates a mental load that slows down your development process. You lose the context of what you were doing, especially if you get distracted by a meeting or a notification.
Finding a specific session after you have closed your terminal is another challenge. The tool saves your history, but a list of directory names and timestamps is not always helpful. You need a system that lets you identify tasks by their purpose rather than their start time. Without clear names for your sessions, you spend more time managing your environment than actually writing code. This problem is common for developers who use AI agents to handle several parts of a project simultaneously.
Resuming sessions with the right context
Claude Code provides a few ways to manage your history through command line flags. The claude --resume command is your primary tool for finding past work. When you run it, the tool displays a list of previous sessions from your current directory. You can see basic details about each session, which helps you pick the right one. This is the built-in way to find a session that you did not finish or one that contains important information you need to reference.
If you are certain you want the very last session you worked on in a folder, you can use the claude --continue flag. This skips the selection menu and drops you straight back into the conversation. This is useful for short breaks, but it does not help you distinguish between different tasks if you are working on multiple things in the same repo. To keep your work organized, you must rely on how you structure your folders and how you interact with the tool from the start.
Using CLAUDE.md to define project identity
The CLAUDE.md file is the best place to define the identity of a session. This file sits in the root of your project and provides the agent with its rules and context. You can use this file to specify the name of the project and the specific goals of the agent. When the session starts, the agent reads these instructions and adopts the persona or focus you have defined. This ensures that every time you open a session in that folder, it feels consistent.
You can also create more specific instructions by placing a CLAUDE.md file in a subdirectory. If you have a folder for your API and another for your frontend, running the claude command from within those folders will use the local instructions. This effectively names the session by its functional area. The agent will focus its Read and Glob tools on that specific part of the tree, keeping the conversation relevant to the task at hand. This is a simple way to create named environments for different parts of your codebase.
Organizing tiles on the adhdTerminal wall
Managing many terminal windows is difficult, but adhdTerminal provides a visual solution. It organizes your Claude Code sessions into tiles on a wall. Each tile represents a distinct session and can be given a specific name, color, and number. This turns a list of text sessions into a visual map of your current workload. You can see at a glance which tile is for your database migration and which one is for your unit tests.
The names you give to these tiles stay the same even if you restart the application. This persistence is a major benefit for developers who struggle with context switching. When you open the app, your workspace is exactly how you left it. The tiles resume their sessions, so you do not have to remember which folder each terminal was in. This visual naming system makes it much easier to track multiple concurrent streams of work without getting overwhelmed by identical-looking terminal windows.
Creating custom agents for specific tasks
You can use subagents to create named specialists within a single session. By placing markdown files in the .claude/agents directory, you define roles that the main agent can call upon. You can name these files based on the task they perform, such as a security auditor or a documentation writer. When you use the /agents command, you can see a list of these specialists and activate them as needed.
This approach allows you to break a complex project into smaller, named sub-tasks. Instead of one long, messy conversation, you have a series of focused interactions. Each subagent has its own instructions, which keeps the context clean and the output relevant. This is a great way to maintain a named structure within your work without needing to open multiple top-level sessions. It gives you a way to organize your thoughts and your agent actions by specific categories.
Managing context with slash commands
As you work, a session can become cluttered with too much information. The context window has a limit, and once it fills up, the agent might lose track of earlier details. You can manage this by using the /compact command. This command asks the agent to summarize the session so far and then clears the detailed history to free up space. It is a way of renaming the history into a concise summary that the agent can still reference.
If you want to keep the same project context but start the conversation over, the /clear command is the right choice. This wipes the current session history while keeping you in the same directory with the same CLAUDE.md instructions. It is useful when you have finished one named task and want to move on to the next without closing the window. Using these commands helps you maintain the identity of your workspace while keeping the technical performance of the tool at its peak.
Tracking costs and usage per session
Naming your sessions also helps you keep track of your resource usage. You can use the /usage and /cost commands to see how many tokens you have used and what the estimated price is for the current session. When you have several different tasks running, checking these stats helps you understand which parts of your project are the most expensive to develop with an AI agent. You can identify if a specific named task is becoming inefficient.
This information is vital for managing your development budget. If a session for a minor bug fix is costing as much as a major feature, you might need to /compact the history or change how you are prompting the agent. By treating each session as a named unit of work with its own cost, you become more disciplined in how you use the tool. It turns the agent into a measurable part of your workflow rather than an unpredictable expense.
Customizing the environment with settings
You can control how your sessions behave by using the .claude/settings.json files. Each project can have its own settings, allowing you to define the rules for that specific environment. For instance, you might set a session to acceptEdits mode for a project where you want the agent to work faster, or plan mode for a project that requires more careful review. These settings are local to the folder, so they define the personality of the session.
If you need to override project settings for your own use, you can use a .claude/settings.local.json file. This file is usually ignored by git, so your personal naming and permission preferences do not affect the rest of your team. For global settings that apply to every session on your machine, you use the ~/.claude/settings.json file. This hierarchy of configuration files gives you the power to control every aspect of your Claude Code environment, ensuring it fits your specific way of working.
Synchronizing work with the mobile app
The adhdTerminal mobile app provides a way to stay connected to your named sessions when you are not at your computer. It mirrors the tile wall from your desktop to your iPhone using an end to end encrypted link. You can see the names and numbers of your tiles and check if an agent has finished a task or is waiting for your permission. This means you do not have to stay glued to your desk to monitor the progress of your work.
If a specific tile needs your attention, you can respond to a prompt or send the next command directly from the mobile app. Because the tiles are named and colored exactly like they are on your desktop, you never lose track of which task you are handling. This continuity makes it much easier to manage a complex workload. You can step away from your desk with the confidence that your named sessions are still working and that you can check on them at any time.
Questions people ask
How do I change the name of a tile in adhdTerminal?
You can edit the tile settings directly within the app interface to change its name, color, or number. These changes are saved locally and will persist every time you open the app. The name you choose helps you identify the specific Claude Code session running in that tile.
Can I have multiple sessions open for the same project?
Yes, you can open multiple sessions in the same directory by running the claude command in different terminal windows or tiles. Each session will have its own independent history and context. Using different names or colors for these tiles in adhdTerminal makes it easy to distinguish between them.
Does the agent know the name I gave to a session?
The agent does not automatically know the name you assigned to a terminal window or a tile. To give the agent a sense of its name or role, you should include that information in the CLAUDE.md file or as a system prompt. This ensures the agent behaves according to the specific task you have labeled.
Read next
- Add build instructions to your CLAUDE.md fileImprove project context by adding specific build commands and environment requirements to your CLAUDE.md file for the AI to follow.
- Add common error fixes to CLAUDE.mdImprove project memory by adding a section for common errors and their fixes to your CLAUDE.md configuration file.
- Add custom shell aliases for Claude CodeSpeed up your workflow by creating shell aliases for common Claude Code invocations and project-specific startup commands.
- Add test commands to a CLAUDE.md fileUnderstand how to define specific test runners and environment flags in your CLAUDE.md file so the tool runs the correct test suite.