Guards

An idle server farm above the prompt: daemon bots collect data from your racks, and your Claude work boosts it

Claude Code's task list as a progress bar above the prompt: a pixel track in the Desktop app, dotted text in the terminal.

Before Claude calls a task done, run your checklist. Fail means Claude is sent back to fix it; repeated failure trips a breaker and hands control back to you.

Run Claude Code subagents in pi or opencode instead of a Claude model. Adds the ext-agent:worker agent type; built-in subagents stay available unless you opt in to blocking them.

A cross-platform Claude Code status line inspired by Trail of Bits' statusline script: model, folder, branch, context bar, cost, time, cache hits, live agents

Dry-runs risky Bash commands (recursive deletes, destructive git, kubectl delete, SQL wipes) and asks Proceed or Cancel first.

Don't reinvent the wheel: finds existing GitHub projects, MCP servers, skills and plugins that fit the conversation and shows them in a small pager above the prompt.

NieR:Automata style panels for messages and tool rows, and a band with Lineland and a usage monitor

A pane listing figures, structures and data files written this session, with the notebook cell or script that made each figure and the data it reads

/tree: the project's directory tree in a side panel, with git change markers, refreshed as files change

A little protein in a side panel that folds a step for every finished run, and misfolds on failures

A pane and toasts for background shells, monitors and subagents: status, elapsed time, latest output line, exit code

Bleach: un Bankai animado al azar cuando una tarea sale bien, colección con /bankai, y hollowificación según el contexto usado

merge 前の codex review を必須化する

Masks credentials in tool output (URL passwords, *_TOKEN / *_SECRET / *_PASSWORD values, API keys, bearer tokens, private keys) before Claude or the transcript sees them. No network, files or processes.

When a Bash command asks for permission, preview its blast radius: rm targets measured (files, size, what git can't restore), plus your own regex -> dry-run rules per command in a chain or pipe. A line under the dialog, details in a side pane. Display only: never changes the decision.

Live subscription usage and pace bars, with an optional provider-neutral delegation advisor.

A weather forecast for your Claude Code usage: context, cost, cache and plan limits above the prompt. 8 languages, any time zone.

Are you moving or hiding in the dark? Watches the session for shipped work (commits, pushes, deploys, passing tests) against edits to tooling, config and docs, and says so in one quiet line when you have been in the dark too long. Zero model tokens, no network, nothing leaves your machine.

gh pr create / pr-push.sh で PR ができたら、その branch の CI を gh-actions-bridge の ws /watch に Monitor で繋ぐ Claude Mod (function hooks)

Guided modernization for any legacy codebase: start with /modernize, get an assessment, an interactive map, the business rules mined from the code, and a plan you approve, then an upgrade, a rewrite or a rebuild with independent proof that the new code behaves like the old. Includes security hardening, a one-page report, specialist agents, and an optional live progress pane (early access).

Picks the reasoning effort for each turn of the main conversation with a System One classifier (hosted Jev or a compatible service) and shows the level on the spinner and on the line that closes the turn. In shadow mode it shows and logs its pick and sends the session's effort; in enforce mode it sends the pick. Requires CLAUDE_CODE_ENABLE_FUNCTION_HOOKS=1.

Brings back /buddy: the small creature that watched you code, above the prompt again. Restores the companion in ~/.claude.json exactly (same species, rarity, hat and stats), hatches an egg for new accounts, levels up from commits and passing tests, unlocks a 49-buddy roster as you ship, reacts to Claude's work for free, answers in a Haiku-written line when you say its name, and shows in a pane...

Harness engineering for Claude Code: attribute every prompt/agent change, keep every candidate, diagnose from traces, hold out a split - and search over Claude Code's own scaffold against your repo's real failures.