More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Termômetro Pro · Claude Code limits

    1

    Stop finding out you hit your limit when it is already over.

    $9

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Termômetro Pro · Claude Code limits

    Termômetro Pro · Claude Code limits

    Example session with this skill installed

    Install Termômetro Pro and tell me how much of my Claude Code limits I have left before I start this refactor.

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

    Done. I installed Termômetro Pro:

    $ python3 install.py
    Backup: ~/.claude/settings.json.termometro-backup-20261108-034000
    Slash command installed: /termometro
    Termômetro status line installed. It appears on your next Claude Code message.
    

    Here is where you stand right now

    ╭─ ◆ MAX 5×  Opus 5.5 ────────────────────────────────────────────────────────────────── termometro ─╮
    │ 5H    ▕██████▋     ▏55%  ≈21 msgs  resets 05:08 │ WEEK    ▕█████▊      ▏48%  resets Mon at 17:20   │
    │ TODAY ▁▃▄▂▆█▅▇ 1.4M normal                      │ CONTEXT 180k  cache 60m  US$ 4.06   → DASHBOARD  │
    ╰────────────────────────────────────────────────────────────────────────────────────────────────────╯
    

    Short answer: you have room for about 21 prompts like your recent ones in this 5-hour block, and at today's pace your week makes it to the reset on Monday.

    How to read it

    PartWhat it means
    ◆ MAX 5× · Opus 5.5Your plan (detected locally) and the model in use
    5H 55% · ≈21 msgsOfficial 5-hour block, and how many prompts like your recent ones still fit
    WEEK 48%Official weekly limit, with the reset time
    TODAYYour tokens today vs your own normal, with the last 8 hours as a sparkline
    CONTEXT · cache · US$Chat size, when the prompt cache expires, and the chat's value at API prices
    → DASHBOARDOpens the offline dashboard (or type /termometro)

    If your pace would exhaust a limit before it resets, the panel turns red and the title says when, for example "▲ 5h block runs out at 15:20". When a busy block renews, you get a desktop notice so you know you can get back to work.

    Before the refactor

    • Long refactors re-read big contexts. If CONTEXT goes past ~250k, start the next step with /clear and a short summary: the "≈ msgs" number goes up right away.
    • Open the dashboard once: the

    plan advisor tells you whether your plan fits your pace, and

    Your week in one card gives you a shareable summary.

    • Optional daily digest: python3 ~/.claude/skills/termometro/scripts/statusline.py --digest 18:00.

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

    About this skill

    Stop finding out you hit your limit when it is already over. Termômetro Pro reads the official limits Claude Code reports and answers the only question that matters mid-task: will I make it to the reset?

    ╭─ ◆ MAX 5×  Opus 5.5 ──────────────────────────────────────────── ▲ 5h block runs out at 03:41 ─╮
    │ 5H    ▕██████████▌ ▏88%  ≈4 msgs ▲ │ WEEK    ▕████████▋   ▏72%  resets Mon at 17:20            │
    │ TODAY ▁▃▄▂▆█▅▇ 1.4M normal         │ CONTEXT 310k → /clear  cache 60m  US$ 4.43   → DASHBOARD  │
    ╰────────────────────────────────────────────────────────────────────────────────────────────────╯
    

    In your status line

    • ≈ prompts left: "≈4 msgs" before the first limit runs out, calibrated on your own history and on what your recent prompts in this chat actually cost.
    • Forecast: "5h block runs out at 15:20" in the panel title, while there is still time to slow down, switch tasks or /clear.
    • Your official 5-hour and weekly limits as clean bars, with reset times and a daily quota until the reset.
    • Your plan badge (PRO, MAX 5×, MAX 20×, TEAM, ENTERPRISE) and the model in use.
    • Context size with a /clear hint, prompt-cache expiry and what the chat is worth at API prices.
    • Color only when something needs attention. Light and dark terminals; narrow terminals get a compact panel, then a mini view.

    Notices that respect your focus

    Never on a timer, only when something happens:

    • At 80% and 95%, and when a limit would run out before it resets.
    • When you can get back to work: the moment a busy 5h block renews ("100% free again") and when your week resets.
    • Optional end-of-day digest at the time you choose.

    The dashboard (one offline HTML file)

    • A weekly thermometer and a weather-style forecast of your week ("Storm on Thursday").
    • Plan advisor: from your official limit history, whether you keep hitting the ceiling ("MAX 5× would have covered your peak at ~20%") or your plan has room to spare.
    • Your week in one card: a 1200×630 PNG to share on X, Reddit or LinkedIn. Totals only, never project or chat names.
    • Official limit history, tokens per day vs your own normal, which chats ate your week, value per model, practical tips.

    Why not just run /usage?

    /usage tells you where you are when you remember to ask. Termômetro tells you, without asking, how many prompts you have left, when you will run out at this pace, and when you are free again. You stay in the CLI and never get surprised mid-task.

    Private by design

    • Python only, 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.
    • One-command install with a backup of your settings; --uninstall restores your previous status line exactly.
    • English and Portuguese. Also reads your Codex CLI limits if you use it.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.

    Requires local Python 3.8. Status line must run once to populate data. Limits only reflect Claude Code CLI usage, not web interface usage.

    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 buying a skill.

    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

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 9 days ago

    Software
    Python

    Permissions

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

    What's inside

    Frequently Asked Questions

    /usage tells you where you are when you remember to ask. Termômetro shows it all the time and goes further: how many prompts still fit, when a limit runs out at your current pace, and a notice the moment you're free again. You never leave the CLI to check.

    It is an estimate, shown with ≈. It divides what is left of your official limit by what your last few prompts in this chat actually cost, using a tokens-per-percent calibration learned from your own history. It shows "calibrating" until it has enough data, usually within your first session.

    No. Everything runs locally from Claude Code's own data. It never sends a request to the model and makes no network calls. The weekly share card contains totals only, never project or chat names.

    Claude Code with a Pro, Max, Team or Enterprise plan and Python 3.8+. Built and tested on Linux (light and dark themes, truecolor and 256 colors, 45 to 200 columns). The macOS and Windows code paths are handled but not yet verified on real machines.

    Yes. The installer backs up your settings first, and python3 install.py --uninstall restores your previous status line exactly. Your history stays in ~/.claude/termometro unless you delete it.

    New in LLM Ops