More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Termômetro Lite · Claude Code limits

    1

    Stop typing /usage or opening the claude.ai settings page in the middle of a task.

    Free

    Security scanned
    Termômetro Lite · Claude Code limits

    Termômetro Lite · Claude Code limits

    Example session with this skill installed

    Install Termômetro Lite so I can see my Claude Code limits in the status line.

    • Read your context and instructions
    • Compiled the term metro lite

    Done. I ran the installer that ships with the skill:

    $ python3 install.py
    Your previous status line was saved to ~/.claude/termometro/previous-statusline.json
    Backup: ~/.claude/settings.json.termometro-backup-20261108-031500
    Termômetro status line installed. It appears on your next Claude Code message.
    

    From your next message on, the bottom of your terminal looks like this:

    ╭─ ◆ MAX 5× ────────────────────────────────────────────────────────────── termometro lite ─╮
    │ 5H    ▕█████▌      ▏46%  resets at 05:03 │ WEEK    ▕████▌       ▏38%  resets Mon at 17:15 │
    │ TODAY 1.4M normal                        │ CONTEXT 92k                                    │
    ╰───────────────────────────────────────────────────────────────────────────────────────────╯
    

    How to read it

    PartWhat it means
    ◆ MAX 5×Your plan, read locally from Claude Code's own account file
    5H / WEEKYour official limits as Claude Code reports them, with the reset time
    Bar colorGreen while calm; amber, orange and red only when a limit needs attention
    TODAYTokens used today vs your own normal (average of your last 14 active days)
    CONTEXTHow big the current chat is; a → /clear hint appears when it gets heavy

    When a limit passes 50%, a discreet line shows what Termômetro Pro would add. In narrow terminals the reset times are hidden so nothing wraps; below ~70 columns it switches to a two-line mini view.

    Next steps

    • Nothing to restart: it shows up on your next message.
    • Light terminal not detected? Set TERMOMETRO_THEME=light.
    • Portuguese? It follows your system language, or set TERMOMETRO_LANG=pt.
    • To undo: python3 install.py --uninstall restores your previous status line exactly.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Stop typing /usage or opening the claude.ai settings page in the middle of a task. Claude Code already knows your plan limits; it just doesn't show them while you work. Termômetro Lite keeps them in view, in one clean panel at the bottom of your terminal, updated as you work. Install it once and you never have to leave the CLI to ask "how much do I have left?" again.

    ╭─ ◆ MAX 5× ────────────────────────────────────────────────────────────── termometro lite ─╮
    │ 5H    ▕█████▌      ▏46%  resets at 05:03 │ WEEK    ▕████▌       ▏38%  resets Mon at 17:15 │
    │ TODAY 1.4M normal                        │ CONTEXT 92k                                    │
    ╰───────────────────────────────────────────────────────────────────────────────────────────╯
    

    What it shows

    • Your official 5-hour and weekly limits: the same numbers Claude Code reports, not estimates, with the reset time.
    • Your plan badge (PRO, MAX 5×, MAX 20×, TEAM, ENTERPRISE), detected locally.
    • How heavy the current chat is, with a /clear hint when the context gets big.
    • Today's tokens against your own normal (your 14-day average), not an arbitrary quota.

    Designed to stay out of your way

    • One framed panel, aligned columns. Color only when something needs attention.
    • Readable on light and dark terminals; narrow terminals get a compact panel, then a two-line mini view.

    Private and light

    • One Python file, zero dependencies. No Node, no network: nothing leaves your machine.
    • Your plan is read from the plan field of ~/.claude.json; credentials, tokens and API keys are never touched.
    • English and Portuguese, following your system language.
    • The installer backs up your settings; python3 install.py --uninstall restores your previous status line exactly.

    Seeing the number is good. Knowing if you'll make it is better.

    Termômetro Pro adds how many prompts still fit ("≈14 msgs"), a forecast that warns you before a limit runs out, a desktop notice the moment a busy block renews, a plan advisor, a weekly share card and an offline dashboard of what ate your week.

    Requires Python 3. Only supports status line integration in Claude Code. Forecasts and API cost tracking are not included in Lite version.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 9 days ago

    • Free with an Agensi account

    Requires

    Python

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Environment Variables
    Listed9 days ago

    What's inside

    Frequently Asked Questions

    Because you only run it after you suspect a problem, and every check pulls you out of the task. Termômetro Lite keeps the same official numbers on screen the whole time, with reset times, and only adds color when a limit actually needs attention. You glance down instead of stopping.

    Claude Code with a Pro, Max, Team or Enterprise plan (the limits come from Claude Code itself) and Python 3.8+. No other dependencies. Built and tested on Linux in light and dark themes and narrow terminals; the macOS and Windows code paths are handled but not yet verified on real machines.

    No. The status line runs locally and reads Claude Code's own data on your machine. It never sends a request to the model and never makes a network call, so it costs zero tokens.

    One command: python3 install.py (or ask your agent to install it). Your current settings are backed up first, and python3 install.py --uninstall restores your previous status line exactly.

    Lite shows where you are: official limits, reset times, today vs your normal and chat size. Pro tells you where you're heading: how many prompts still fit, a forecast before a limit runs out, a notice the moment a busy block renews, a plan advisor, a weekly share card and an offline dashboard.

    New in LLM Ops