Status line

各ターンの終わりにコンテキスト使用率を天気アイコンで表示する

Live pane with every step of the current plan, the current step marked, and a done/total count

See Claude's background shell commands: live output in a pane, status, and a toast when one fails

Counts down the 60-minute prompt cache TTL since the last model response

Replace protected people's names, Taiwan ID numbers, phone numbers and emails with placeholders before the model reads them.

Raily is a personal AI agent that finds people worth meeting for you (friends, dates, business contacts) and handles the first contact. This mod shows your Raily inbox above the Claude Code prompt: new matches with name, city and score, contact requests, unread chats and questions from your agent, with a toast for each new event. Read-only. Needs a Raily account (18+) and the Raily MCP server...

Claw mod: wakes Claw up on a schedule inside a running Claude Code session. Daily briefing at a set time, periodic checks for overdue tasks, quiet hours, and an optional auto-run mode that makes Claude do the heartbeat by itself.

Claw mod: talk to Claw from your phone. Messages from your own Telegram chat start a turn, the answer is sent back, briefings and check-ins are pushed to you, and /approve on your phone unlocks one action that Claw Guard is holding.

Keeps rate-limit usage in the status line, warns before agent fan-out near a limit, and blocks it at the edge.

Pane listing the Claude Code and Codex sessions working in this repo or folder

Before/after simulator screenshots on the same step as a UI edit, side by side in a pane, with a changed-pixel percentage and pinned baselines

Wraps web, MCP and third-party tool output as untrusted data and defangs prompt-injection lines before the model reads them

Attaches fresh React Native redboxes and native crash logs from running simulators and emulators to your prompt, and warns when an edit needs a native rebuild

Type-checks and lints the file Claude just edited and puts the errors in that same tool result

Two-way secret scrubbing: redacts credentials before the model sees them and blocks writing them into tracked files or shell commands

Attaches a file's unresolved GitHub PR review threads to the model's Read and Edit results, so review comments get fixed in the file it is already touching

Shows the folder Claude is working in on the status line.

A RollerCoaster Tycoon-style park that Claude builds while it works: tool calls lay track, coasters get test runs, crowds pour in, and parks compete on a global leaderboard.

Claude's questions drop from your Mac's notch as a native card. Pick with 1-4, type your own answer, or press esc to answer in Claude Code as usual.

Guards the 5h usage window: caps expensive-tier spawns per turn, keeps Fable spawns manual-only, draws a colour-coded usage gauge above the prompt (5h bar, reset time, turn burn, cache warmth, context size), and offers handoff + clear past a context size.

Confirm dialog before outward-facing sends and out-of-vault deletes; off switch for unattended runs; always-on audit log.

Hands the work to a fresh session before a long context degrades the model, and carries on there.

Holds risky shell commands (rm, git reset/clean/force-push, kubectl, terraform, docker, SQL and more), shows what they would change and whether you can undo it, and runs them only on your say-so.

Shows when the prompt cache lapses, keeps it warm with a few pings while idle, then compacts once and waits