All mods

Sends a desktop notification when a long turn finishes, so you can look away while Claude works.

Blocks Write, Edit and MultiEdit calls that target a path outside the project root or inside .git, plus optional deny globs.

A one-row powerline-style band above the prompt: directory, git branch and state, model, context fill.

Blocks Write, Edit and Bash calls that would put an API key, token or private key into a file or command.

Blocks Read, Write, Edit, Grep, Glob and Bash calls that touch .env files, private keys and credential stores.

Adds elapsed time, the running tool and the tool-call count to Claude Code's own spinner.

Blocks edits that make tests pass by weakening them: new skip/only markers, fewer assertions, a gutted test file, or deleting a test file.

A countdown in the prompt footer: under the hint line on the terminal, left of the model picker on the desktop: time left on the prompt cache and how much of the last request it served. Olive while there is plenty, warmer at 10, 5 and 2 minutes left (scaled down for a 5-minute cache). The lifetime follows Claude Code's own rules: 1 hour on a subscription, 5 minutes on an API key or usage credits.

Debajo de cada acción de Claude, una línea en español normal que explica qué está haciendo (leer, crear, mover, borrar, buscar...). Para entender a Claude sin saber informática.

/mejora y una frase: Claude te hace 3-5 preguntas, escribe un prompt con contexto (sin rol), lo puntúa y lo corrige hasta el 9/10, con bucle o sin bucle. Método de diariovida.com.

Modo grabación: /grabar oculta en pantalla correos, rutas de usuario, IPs, claves y los nombres que tú elijas, para grabar sin enseñar datos

Cuando Claude va a borrar algo, en vez de borrarlo lo manda a una papelera (~/.papelera-claude) para que puedas recuperarlo. /papelera te enseña qué hay dentro.

Tu Claude se ve distinto: una franja encima del cuadro de texto con tus reglas activas y el «pensando» en español castizo.

Picks the model and the reasoning effort for each task with TypeSafe's Jev, a System One decision model reached either through TypeSafe's own API or the Vercel AI Gateway. Routes subagent models at agent.spawn and the main loop's effort and model at turn.step, in both directions, falling back to the engine's built-in classifier when no key is set.

Bottom-left retro RPG companion: a pet fed by cache hits and commits, failing test suites as bosses.

Bottom-grid row: context fill and prompt cache (hit rate, countdown from the TTL the API reports). Toasts when context is full or the cache expired.

Band showing avg time to first token and output tokens/s

Band above the prompt: current model and reasoning effort.

Band showing rtk gain: tokens saved and quota preserved

Bottom-grid row: plan and session/week usage. Toasts when limits near.

A pixel-art monster in a side pane that eats tokens and shows context, limits and activity at a glance

Claude Clubhouse: a toolbar above the prompt, rooms for usage, agents, commands, tool rules and colors, and whole-window themes on macOS

Buttons above the prompt: retell recent turns in plain words, or open what a background check flagged that you might miss. Results go to Claude Code's own pane; nothing enters the transcript or the model's context.

An orange cat above the prompt that reacts to your commits, tests and builds