Guards

Draw a formula and show the PNG in a pane

A small animated pet above the prompt that reacts to what Claude is doing

At the end of every turn Claude prints a receipt: files touched, commands run, tokens and what the turn cost

When a tool call fails, the spinner dies: the line goes grey and WASTED fades in, in hand-drawn pixel lettering

A badge under each Bash tool row naming the environment the command targets (AWS profile, kube context, ...). Rules, tiers, format and label scripts are configured in settings.json pluginConfigs. Display only: never touches tool output or model context.

Enforces git identity, attribution and branch-naming rules on Bash calls

Stops Claude adding banner, multi-line, long or narrating code comments

Shows the OpenSpec changes of the open project under the prompt, for local roots and stores.

What Claude is building (website, mobile app, desktop client, Unity game…), live in a panel of the client (desktop, mobile in Remote Control, terminal): Claude hooks up the project and the network bridge itself, and drives the same target as you

Clawd's HQ: Clawd lives in a pixel art office, mimes what Claude does, and a game HUD shows the model, context and rate limits

Thumbnails in the terminal of the images Claude reads or shares (Read, SendUserFile, MCP screenshots), and a /gallery view on every screen

Evita que el contexto se infle: frena lecturas de archivos enormes y recorta salidas gigantes de comandos

Oculta a Claude los datos personales (DNI, NIE, IBAN, teléfonos, emails, claves) cambiándolos por falsos válidos y los restaura al ejecutar, todo en tu máquina

Progreso del plan sobre el prompt: fase actual, barra por fases, subagentes, tiempo transcurrido y estimación de lo que falta

Los subagentes usan Sonnet por defecto; Claude recibe la regla para elegir Opus o Fable cuando toca

Live progress bars above the prompt for background shells and subagents

Opens a card when Claude's final answer omits a recorded file change or a recognized command with side effects, or makes a test or Git claim the recorded results do not fully back. Step through the diffs and prepare an editable question.

Full-duplex voice conversation with your running Claude Code session, powered by OpenAI gpt-live-1. Toggle with /talk.

One plain-language line above the prompt saying roughly where a long task is, written by a cheap fast model.

記下每一版計畫和你退回時說的話，核對你加的需求有沒有寫進最新版，計畫通過後再核對做完沒

提示框上方顯示還沒推出去的 git 狀態：分支、未 commit 檔案、未推 commit、已合併未刪的分支

コンテキスト残量とプランの使用量をステータス行に常時表示する

Keeps the last part of your 5-hour and weekly quota for you. At the reserve, spare10 holds all work and asks you whether to continue. After a Resume, it asks again at the floor, by default at 5% left. Shortly before the reset, it lets work use the reserve.

A pixel slime above the prompt that grows as you work with Claude