ToBeBestMe

A mod that runs inside Claude Code itself. Not a separate app, service, website, command-line tool or account: it changes Claude Code's own interface, with a pane beside the chat, a line above the prompt, a tool Claude answers from, and alerts through Claude Code's own channel. There is no server, no sign-up and no API key, and it runs on your computer.

What it shows: what's waiting for you, what your week has left, and what every workflow is doing, across all your Claude Code projects. It is for people who keep several projects going at once, on a Pro or Max plan or an API key, and want to stop losing track of which chat asked a question, when a limit will run out, and what a workflow is about to cost.

/plugin install tobebestme --marketplace vradulthub-cloud/tobebestme

Landing page · What it reads and sends · Why it exists · Changelog

ToBeBestMe is an independent open-source project, not made by or affiliated with Anthropic. Claude and Claude Code are trademarks of Anthropic, PBC.

It tracks the three things every chat spends, and none of them is in view while you work:

• Usage limits: the weekly limit on Claude plans, and a model's own weekly limit where there is one.
• Money: what each request costs on an API key or team billing.
• Context: the conversation Claude re-reads on every turn. As it grows, each turn costs more.

What you get

Type /tobebestme to open the pane. It has six tabs (Home, Workflows, This session, Usage, Skills and Settings); keys 1 to 6 switch between them in the terminal. Screenshots show sample data.

Home: the week, and what needs you

Home reads Claude Code's own records for every project on the computer, including chats that never ran ToBeBestMe.

• One drawing of the week. The weekly figure and a forecast sentence, each day's work by project, the line of use so far, the dashed line to where the limit runs out, and the hours you would be without Claude before it resets. Press a day or a project to pick it out; the rest step back and a line says what it held.
• Needs you. Every chat that stopped for you in the last week, most pressing first: a permission prompt, a question, a plan to approve, a usage limit, a pull request waiting. The most urgent is a card of its own. In the desktop app a suggested reply shows the whole message and asks before it sends; Reply… takes your own words. Elsewhere, Copy resume command carries the chat on in a terminal.
• Pick up where you left off. The latest chat in each project, with its last message, branch, pull request and loose ends such as 3 files changed, not committed.
• What stands out. Up to three plain observations. A saving is drawn before you act on it: the week as run, against the week with the idea applied.
• Pull requests, from GitHub. When the GitHub CLI (gh) is installed and signed in, your open pull requests that need you join Needs you: review requested, changes requested, checks failing, merge conflicts, and ready to merge, each with a button to open it and one to copy its link. They are read in the background at most every five minutes. With no gh, or not signed in, nothing is shown and Settings › What ToBeBestMe reads says why.
• Projects, open chats and running workflows.

The weekly report card

On Usage, the week as one image to post: the weekly figure, each day by project, days used and your streak, top projects, models, workflows and savings. Save report card writes an SVG to ~/.claude/tobebestme/reports/ (and a PNG when rsvg-convert, ImageMagick, Inkscape or macOS Quick Look is there) and copies a few lines of text to paste with it. Hide project names swaps them for Project 1, 2, 3 first. Nothing is uploaded.

On your phone

Attach a phone and the mod draws a short lobby in the mobile app: what is waiting for you (with the replies the chat itself suggested, each confirmed before it is sent), the limit that binds first and when it runs out, and what is running. The app has no text field yet, so replies are the suggested buttons.

Workflows

A workflow is a script Claude writes to split a big job across many agents. One run can start hundreds of agents and go for hours.

• Before it starts: what the run is, its stages, how many agents it starts, a likely cost range from your own finished runs, and anything worth changing. With Check each workflow before it runs set to Ask me first, Claude waits for your answer in the chat: Run workflow, Run every agent on Sonnet (every agent, including stages the script put on a stronger model) or Don’t run.
• While it runs: every agent and what it is doing this moment, its model and cost. Pause new agents, Agents at once, Model for new agents and Stop workflow… act on agents that have not started; finished work is kept.
• Afterwards: the costliest runs, and ways to spend less, each with what it would have saved on your last 30 days of runs.

Make it a rule… shows the line it would add to your user CLAUDE.md, as it is now and as it will be, and writes it when you press Save change.

This is the beginning of the README. Read the rest at the original. Read the full README ↗