Automation

Agent Plugins、Claude Code、Codex向けの品質支援プラグイン。コードと文書を扱う作業、計画やバグ調査の指針を示し、フックで入力と操作を確かめる。

A weekly check of your Claude and Codex plugins, global npm packages and adapted skills for newer versions, with a one-line summary of each

Turn global skills, rules, agents and tools on or off per repo with named profiles

Report model drift and deny only positively invalid stale-pin launches.

Corrects the model tier of every subagent spawn (agent.spawn) with a live Jev judgment (haiku / sonnet / opus), instead of relying on /delegate's static routing table. A risky task (touches production, money, or irreversible data) is always forced to opus regardless of confidence.

Saves how long every turn took as a row in the chat, logs every prompt's timing, and adds /timings to look back.

Hides SecChain's secret values in prompts, tool results and instructions before the model reads them, through secchain mask.

Snapshots the working tree into hidden git refs at the start of every turn, with /checkpoints to list them and /undo-turn to roll back.

Stops Claude from editing files or running mutating git commands on main or master, and points it at a worktree instead.

Automated multi-agent code review cycle. Spawns reviewers in parallel (plus Codex when available), applies fixes per CLAUDE.md policy, admits a commit only when a reviewer saw exactly what it records, and asks before a push, pull request, merge or release you did not request, unless you let it go further.

agentctl Mod: agentctl CLI (tools/agentctl) が ~/.agentctl/claude/ /inbox/ に置いたメッセージを $.prompt.submit で Claude に届け、中断の依頼を $.turn.abort で行い、結果を acks/ に書く。付属スキル agentctl は、agent に agentctl の使い方を教える。

Counter-Strike 1.6 radio calls for Claude Code: 'Fire in the hole' on deploys, 'Bomb has been defused' when a long turn lands. /radio to toggle.

Cross-model review and delegation: /mm:review and /mm:run, built on mmrun

A per-project progress board: the main model records each conversation's work as nodes through the progress tool, kept under ~/.claude/progress/ or committed with the project; an optional Artifact canvas mirrors it

Tallies the reminders Claude Code injects for the model, per session; drops the token counter and repeated attribution blocks; /reminders prints the tally

A small Claude Code plugin with guards, model and effort rules, compaction and cache notes, and role agents.

Switches Claude Code's todo tools on for models that have them left out

Opt-in: after a long turn, hear one spoken sentence about what Claude did.

Plain English under every permission prompt: what the command does and how risky it is.

A one-line TL;DR under long answers, written by haiku.

When Claude re-reads a file that changed since it last read it, it gets only the changed hunks (current line numbers, a few context lines) instead of the whole file.

Huge Bash/PowerShell outputs without the context bill: Claude gets the head and tail, the full text goes to a file it can open.

Stops Claude from re-reading a file it already read in full and that has not changed since; counts the tokens saved.

Logs every function-hook event (on '*') to JSONL for inspection.